Perspectives on Bullying : Research on Childhood, Workplace, and Cyberbullying [1 ed.] 9780826194633, 9780826194626

Citation preview

Perspectives on Bullying Research on Childhood, Workplace, and Cyberbullying Roland D. Maiuro Editor

Copyright © 2015 Springer Publishing Company, LLC All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of Springer Publishing Company, LLC, or authorization through payment of the appropriate fees to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400, fax 978-646-8600, [email protected] or on the Web at www.copyright.com. Springer Publishing Company, LLC 11 West 42nd Street New York, NY 10036 www.springerpub.com ISBN: 978-0-8261-9462-6 e-book ISBN: 978-0-8261-9463-3 15 16 17 18 / 5 4 3 2 1 Content herein was selected from Violence and Victims (ISSN: 08866708; eISSN: 19457073) published bimonthly by © Springer Publishing Company, LLC. For more information, visit www.springerpub.com/journals/violenceand-victims.html The author and the publisher of this Work have made every effort to use sources believed to be reliable to provide information that is accurate and compatible with the standards generally accepted at the time of publication. The author and publisher shall not be liable for any special, consequential, or exemplary damages resulting, in whole or in part, from the readers’ use of, or reliance on, the information contained in this book. The publisher has no responsibility for the persistence or accuracy of URLs for external or third-party Internet websites referred to in this publication and does not guarantee that any content on such websites is, or will remain, accurate or appropriate. Special discounts on bulk quantities of our books are available to corporations, professional associations, pharmaceutical companies, health care organizations, and other qualifying groups. If you are interested in a custom book, including chapters from more than one of our titles, we can provide that service as well. For details, please contact: Special Sales Department, Springer Publishing Company, LLC 11 West 42nd Street, 15th Floor, New York, NY 10036-8002 Phone: 877-687-7476 or 212-431-4370; Fax: 212-941-7842 E-mail: [email protected] Printed in the United States of America.

Contents

Introduction  v Chapter 1:  Gender, Bullying Victimization, and Education  1 Ann Marie Popp, Anthony A. Peguero, Kristin R. Day and Lindsay L. Kahle Chapter 2: Short-Term Stability and Prospective Correlates of Bullying in Middle-School Students: An Examination of Potential Demographic, Psychosocial, and Environmental Influences  15 Dorothy L. Espelage, Kris Bosworth and Thomas R. Simon Chapter 3: Understanding Ecological Factors Associated With Bullying Across the Elementary to Middle School Transition in the United States  31 Dorothy L. Espelage, Jun Sung Hong, Mrinalini A. Rao and Robert Thornberg Chapter 4: Individual and Social Network Predictors of Physical Bullying: A Longitudinal Study of Taiwanese Early Adolescents  49 Hsi-Sheng Wei and Wonjae Lee Chapter 5: Psychometric Properties of the Cyberbullying Questionnaire (CBQ) Among Mexican Adolescents  65 Manuel Gámez-Guadix, Fabiola Villa-George and Esther Calvete Chapter 6: Do Networking Activities Outside of the Classroom Protect Students Against Being Bullied? A Field Study With Students in Secondary School Settings in Germany  81 Gerhard Blickle, James A. Meurs and Christine Schoepe Chapter 7: The Differential Impacts of Episodic, Chronic, and Cumulative Physical Bullying and Cyberbullying: The Effects of Victimization on the School Experiences, Social Support, and Mental Health of Rural Adolescents  99 Paul R. Smokowski, Caroline B. R. Evans and Katie L. Cotter Chapter 8: Teachers Bullied by Students: Forms of Bullying and Perpetrator Characteristics  117 Teemu Kauppi and Maili Pörhölä

ivContents

Chapter 9: Perpetrators and Targets of Bullying at Work: Role Stress and Individual Differences  135 Stig Berge Matthiesen and Ståle Einarsen Chapter 10:  Workplace Bullying, Emotions, and Outcomes  155 Lars Glasø and Guy Notelaers

Introduction

T

he prospect of elevating one’s social status is a common motivating factor for bullying in college, just as it is in elementary school... A recent survey found that 35 ­percent of the U.S. workforce report being bullied at work… A majority of American school districts have no policies protecting LGBT students from bullying. Once thought to be solely the bane of children in school settings, bullying is now understood among social scientists to affect adolescents and adults in the workplace and the cyber world—often resulting in substantial, long-term emotional and physical harm. This collection of prominent research articles published in the peer-reviewed journal Violence and Victims focuses much-needed attention on a serious problem that has been long ignored and is now on the rise. Bullying experts from a variety of disciplines—psychology, psychiatry, sociology, criminology, counseling, and social work—provide comprehensive, interdisciplinary coverage of bullying in school settings, adulthood, the workplace, and online. They present current research related to predictive factors for bullying, perpetrators of bullying, victimization, and prevention programs. These articles have been selected on the basis of those most frequently downloaded from the online editions of Violence and Victims. This diverse collection of writings opens with “Gender, Bullying Victimization, and Education,” which focuses on gender in regard to the type of bullying experienced and its link to educational outcomes. Bullying has traditionally been associated with early adolescence, a topic addressed in both “Short-Term Stability and Prospective Correlates of Bullying in Middle-School Students: An Examination of Potential Demographic, Psychosocial, and Environmental Influences” and “Understanding Ecological Factors Associated With Bullying Across the Elementary to Middle School Transition in the United States.” Since early adolescent bullying is a worldwide problem, “Individual and Social Network Predictors of Physical Bullying: A Longitudinal Study of Taiwanese Early Adolescents” discusses the relationship between the perpetrators of bullying, their economic status, and family life. Cyberbullying has segued from the adolescent to the adult population and is on the rise. This is particularly problematic since legislation regarding cyberbullying has not kept pace with its increase, leaving many law enforcement personnel at a loss as to how to best handle it. “Psychometric Properties of the Cyberbullying Questionnaire Among Mexican Adolescents” demonstrates the use of this questionnaire, an instrument for measuring the perpetration and victimization of adolescent bullying via new technology along with an analysis of gender differences within cyberbullying. “Do Networking Activities Outside of the Classroom Protect Students Against Being

viIntroduction

Bullied? A Field Study With Students in Secondary School Settings in Germany,” discusses the effect of external network activities on reducing bullying within a classroom setting that is conducive to bullying. “The Differential Impacts of Episodic, Chronic, and Cumulative Physical Bullying and Cyberbullying: The Effects of Victimization on the School Experiences, Social Support, and Mental Health of Rural Adolescents,” examines the impacts of past, current, and chronic physical bullying and cyberbullying on youth in rural settings. Bullying of adults occurs with a surprising frequency. “Teachers Bullied by Students: Forms of Bullying and Perpetrator Characteristics,” surveyed 70 teachers who were bullied by students and examines the form of bullying they received, the characteristics of students who bully, and how the students who bully behave in their peer relationships. Workplace bullying has always existed but was not closely examined as a phenomenon in the past. “Perpetrators and Targets of Bullying at Work: Role Stress and Individual Differences” and “Workplace Bullying, Emotions, and Outcomes” both discuss workplace bullying and the adverse effects it has on employees. Perpetrators of workplace bullying are often very aggressive and its victims include those with low self-esteem and poor social competency. Workplace bullying can affect not only individual employees who are bullied but can have adverse effects on the workplace as a whole. Whether someone is a perpetrator, victim, or bystander, bullying can have an adverse impact on people of every gender, nationality, and age, and can occur in any number of settings. This collection of scholarly articles from prominent bullying experts worldwide discusses the most current understanding and thinking about bullying that will hopefully lead to solutions toward its reduction.

Violence and Victims, Volume 29, Number 5, 2014

Gender, Bullying Victimization, and Education Ann Marie Popp, PhD Duquesne University, Pittsburgh, Pennsylvania

Anthony A. Peguero, PhD Virginia Tech, Blacksburg, Virginia

Kristin R. Day Miami University

Lindsay L. Kahle, MA Virginia Tech, Blacksburg, Virginia School bullying has detrimental consequences for its victims, including undermining students’ educational outcomes. Furthermore, gender has been shown to play a significant role in determining the type of bullying victimization experienced and educational outcomes. This research examines whether an interaction between gender and bullying victimization exists as well as its impact on educational outcomes (i.e., academic selfefficacy and educational achievement). Multivariate regression analyses, drawing on the Educational Longitudinal Study of 2002, reveal that the interaction between gender and bullying victimization is linked to disparate educational outcomes. The findings and their implications are discussed regarding understanding the relationship between gender, bullying victimization, and education.

Keywords: bullying; gender; youth violence; schools

B

ullying victimization has garnered more attention in recent years because of the change in the perception of bullying as a “rite of passage” to a serious social problem that has lasting, negative consequences for its victims (Espelage & Swearer, 2010; Hong & Espelage, 2012; Peguero, 2012). Most of the research on the consequences of bullying victimization has focused on the psychological and social effects on the individual; however, limited research has explored the association between bullying victimization and educational outcomes. Some studies have found that bullying victimization is associated with diminished educational outcomes (Hanson, Austin, & Zheng, 2010; Juvonen, Wang, & Espinoza, 2011; Nakamoto & Schwartz, 2010). In addition, research also recognizes gender differences in bullying victimization (Faris & Felmlee 2011; Klein, 2012; Wilcox, Tillyer, & Fisher, 2009). For instance, there is some research that suggests that girls are more likely to be victims of verbal aggression, insults, gossip, manipulation, and social isolation by their peers, whereas boys are more likely to be victims of ­physical aggression (Dukes, Stein, & Zane, 2010; Finkelhor, 2008; Popp & Peguero, 2011);

© 2014 Springer Publishing Company1

2

Popp et al.

h­ owever, there is also research suggesting that, for girls, there is an increasing trend of engaging in physical bullying (Espelage & Swearer, 2003; Klein, 2012; Swearer, 2008). Although the research literature reveals a gendered pattern in the type of bullying victimization, it is not clear whether the consequences of being bullied are, themselves, gendered in terms of ­educational outcomes. Given that educational attainment lays the foundation for adulthood—not just in terms of economic success but across several psychological, social, and physical d­ imensions—it is imperative to explore factors that are associated with negative educational outcomes. Two measures of educational outcomes, academic self-efficacy and educational achievement, have received considerable amounts of attention over the past decade (Caprara, Vecchione, Alessandri, Gerbino, & Barbaranelli, 2011; Kao & Thompson, 2003; Kingston, Hubbard, Lapp, Schroeder, & Wilson, 2003). Academic self-efficacy reflects the student’s level of confidence or belief that he or she can successfully accomplish educational assignments and tasks (Bandura, 1977; Caprara et al., 2011; Pajares, 2008). Educational achievement is typically measured as a student’s standardized test scores (Buchmann, Condron, & Roscigno, 2010; Kao & Thompson, 2003). The research literature suggests that there are extensive gender differences in academic self-efficacy and educational achievement. For example, as girls progress through the school system, their academic self-efficacy diminishes (Huang, 2013; Pajares, 2008) and there is an established gender disparities in test scores, which has remained relatively consistent over time (Buchmann, DiPrete, & McDaniel, 2008). The important question that remains is why does this gender difference in educational outcomes exist? Because bullying has serious negative consequences for its victims and the bullying experience is gendered, it is possible that gender disparities in educational outcomes may be related to this type of school-based victimization. Further research exploring the relationships between gender, bullying victimization, and educational outcomes (i.e., academic self-efficacy and educational achievement) is warranted. This study extends the literature on gender, bullying victimization, and educational outcomes in the following ways. First, by examining two different educational outcome measures: academic self-efficacy and educational achievement. Second, this study explores the relationship between direct and indirect bullying and educational outcomes, extending current literature, because much of the previous focus has been on the direct bullying rather than indirect bullying. Finally, this study examines whether or not gender moderates bullying victimization and academic self-efficacy and educational achievement. To explore these issues, this study uses data from the Educational Longitudinal Study of 2002, a nationally represented stratified sample of 10,440 10th-grade public school students, and employs regression analyses.

LITERATURE REVIEW Bullying Victimization Typically, bullying victimization is defined as a systematic and reoccurring type of aggression by more powerful peers toward a weaker individual (Espelage & Swearer, 2010; Hong & Espelage, 2012; Olweus, 1993). The research emphasizes two broad categories of bullying victimization: direct and indirect. Direct bullying victimization is physical aggression and harassment, such as hitting, pushing, kicking, and the destruction of

Gender, Bullying, and Education3

property (Espelage & Swearer, 2010; Hong & Espelage, 2012; Olweus, 1993). Indirect bullying victimization is described as verbal aggression, gossiping, manipulation, and social ­isolation of the victim, which is intended to damage the victim’s social status and self-esteem (Dukes et al., 2010; Klein, 2012; Swearer, 2008). Studies also reveal that gender plays a significant role in the type of bullying victimization a student experiences. Girls are more likely to experience indirect forms, whereas boys are more likely to experience direct forms of bullying victimization (Dukes et al., 2010; Finkelhor, 2008; Popp & Peguero, 2011); on the other hand, there are findings indicating that girls are ­increasingly getting into physical forms of bullying (Espelage & Swearer, 2003; Klein, 2012; Swearer, 2008). The consequences of bullying victimization tend to be serious, negative, and long-­ lasting regardless of type of bullying. Bullying victimizations have commonly been associated with psychological and behavioral problems (Espelage & Swearer, 2010; Finkelhor, 2008). In addition, research also finds that girls experience greater psychological distress than boys as a result of experiencing indirect bullying, which may be caused by the heightened importance girls place on the social group (Faris & Felmlee, 2011; Klein, 2012; Young, Boye, & Nelson, 2006). Although there is substantial research exploring the psychological and social consequences of bullying victimization, limited research has addressed the effect of bullying victimization on educational outcomes, especially academic self-efficacy (most of the existing research has focused on educational achievement). This is surprising because bullying victimization among adolescents frequently takes place within the school context (Espelage & Swearer, 2010; Hong & Espelage, 2012; Peguero, 2012) and it is, therefore, reasonable to expect that it will have a negative effect on the student’s academic success. Research demonstrates that bullying victimization is associated with lower levels of educational achievement (Juvonen, 2011; Juvonen et al., 2011; Nakamoto & Schwartz, 2010) and that such detrimental effects may be mediated by resulting psychological and adjustment problems (Schwartz, Gorman, Nakamoto, & Toblin, 2005; Wei & Williams, 2004). To date, research has not found gender disparities among the effects of bullying victimization and educational outcomes. However, Wei and Williams (2004) argued that their model works equally well for both boys and girls, but they did not distinguish between direct and indirect forms of bullying.

Gender and Education: Academic Self-Efficacy and Educational Achievement Although there are no gender differences in innate intellectual potential, gender disparities exist in educational outcomes and are the result of external factors such as societal expectations and student experiences in school (Buchmann et al., 2008; Morris, 2012). To understand and improve student performance, gender disparities in educational outcomes have remained a central topic of empirical analyses, especially regarding gender disparities among academic self-efficacy and educational achievement (Buchmann et al., 2008; Huang, 2013; Morris, 2012; Pajares, 2008). From the onset of formal education, boys are at higher risk of delayed entry into kindergarten (Malone, West, Denton, & Park, 2006). Overall, there is substantial evidence that girls academically outperform boys in reading, grades, and test scores (Buchmann et al., 2008; Tach & Farkas, 2006). Academic self-efficacy is defined as an individual’s expectations, convictions, confidence, and beliefs about what he or she can accomplish in various educational situations

4

Popp et al.

(Bandura, Barbaranelli, Caprara, & Pastorelli, 2001; Caprara et al., 2011; Pajares, 2008). The perceptions that students hold about themselves, in relationship to their academic competence, have considerable impact on what the students do with the knowledge and skills they possess. Academic self-efficacy impacts a student’s decisions about effort, determination, and his or her belief in the ability to effectively accomplish the assigned academic tasks (Caprara et al., 2011; Urdan & Pajares, 2006). Higher levels of student self-efficacy are, in turn, associated with increases in educational achievement and attainment (Caprara et al., 2011; Pajares, 2008; Urdan & Pajares, 2006). During elementary school, girls and boys report similar levels of confidence about their ability to achieve educational success; however, by middle school, girls’ academic self-efficacy begins to diminish (Huang, 2013; Pajares, 2008). Because academic selfefficacy is influenced by school-based interactions and relationships (Pajares, 2008; Urdan & Pajares, 2006), it is possible that bullying victimization could be associated with the academic self-efficacy, and the relationship may differ by gender because boys and girls experience different types of bullying. Although educational achievement historically referred to students’ grades or grade point average, contemporary research has measured educational achievement primarily using standardized test scores (Buchmann et al., 2008; Morris, 2012). Research has found a positive relationship between the influence of educational achievement on the student’s future educational attainment, employment, physical health, and psychological well-being (Buchmann et al., 2008; Morris, 2012). Furthermore, because educational achievement is an important indicator of future success, and gender disparities exist in these outcomes, researchers are concerned with identifying the factors that produce gender disparities in educational achievement (Buchmann et al., 2008). Gender disparities in educational achievement have been attributed to many factors, including gender disparities among teacher expectations; differential parenting styles; and race, ethnicity, and socioeconomic status (Brown & Iyengar, 2008; Entwisle, Alexander, & Olson, 2007; Wood, Kaplan, & McLoyd 2007). What remains unresolved, however, is the mechanism through which gender operates to produce different outcomes in academic self-efficacy and educational achievement. Furthermore, because there are established gender differences in educational outcomes and bullying victimization (Buchmann et al., 2008; Klein, 2012; Peguero, 2012), it is probable that the effects of bullying victimization on educational outcomes is moderated by the ­student’s gender.

Additional Factors Associated With Bullying and Education Research points to various student and school factors that are related to bullying and education (i.e., academic self-efficacy and educational achievement). For student characteristics, studies suggest that race and ethnicity affect the likelihood of bullying victimization and educational inequality (Bradshaw, Waasdorp, Goldweber, & Johnson, 2013; Kao & Thompson, 2003; Peguero, 2012). Recent research indicates that family socioeconomic status is correlated with being a victim of bullying, academic self-efficacy, and educational achievement (Brown & Iyengar, 2008; Peguero & Williams, 2011; Urdan & Pajares, 2006). In general, research suggests that schools that are larger, with higher levels of poverty, and in urban locations typically have increased levels of violence, bullying victimization, and barriers to educational progress for students (Bradshaw et al., 2013; Hong & Espelage, 2012; Morris, 2012; Wood et al., 2007).

Gender, Bullying, and Education5

THE CURRENT STUDY: GENDER, BULLYING VICTIMIZATION, AND EDUCATIONAL OUTCOMES In summary, this study explores the relationship between gender, bullying victimization, and educational outcomes. This study extends the existing literature by exploring the effect of bullying victimization on educational outcomes and investigating whether gender interacts with bullying victimization to create gender-specific effects on educational outcomes.

METHOD Data Source Data for this research is drawn from the Educational Longitudinal Study of 2002 (ELS). ELS is a longitudinal survey administered by the Research Triangle Institute (RTI) for the National Center for Education Statistics (NCES) of the U.S. Department of Education. ELS is designed to monitor the transition of a nationally representative sample of young people as they progress from 10th grade through high school and on to postsecondary education and/or the world of work (NCES, 2004). These data include information about the experiences and backgrounds of students, parents, and teachers and descriptions of the schools the students attended. The sample for these analyses included 5,320 female and 5,120 male public school students who are White American, Black/African American, Latina/Latino American, and Asian American. Racial and ethnic minority groups are oversampled in ELS to obtain a sufficient representation for statistical analyses. In turn, the sample weights used in the analyses are calculated by NCES to compensate for the sampling design and for nonresponse bias. Educational Outcomes: Academic Self-Efficacy and Educational Achievement. To measure academic self-efficacy, students were asked to describe their understanding and mastery of educational material during the first semester or term of the 2001–2002 school year. This measure was constructed from student reports in which they describe themselves as being confident on (a) doing an excellent job on math tests, (b) understanding the most difficult material presented in math texts, (c) understanding the most complex material presented by my math teacher, (d) mastering the skills being taught in math class, (e) doing an excellent job on math assignments, (f) doing an excellent job on English tests, (g) understanding the most difficult material presented in English texts, (h) understanding the most complex material presented by my English teacher, (i) mastering the skills being taught in English class, and (j) doing an excellent job on English assignments. Reliability of the scale was determined using Cronbach’s alpha, which is a measure of internal consistency, and the Cronbach’s alpha for academic self-efficacy is .92. Educational achievement is measured by using the standardized measure preconstructed by RTI and NCES. ELS includes a reading and math composite score based on standardized tests developed by the Educational Testing Service (ETS) in math and reading. The composite score is the average of the math and reading standardized scores, restandardized to a national mean of 50.0 and standard deviation of 10 (see NCES, 2004 for further detail). Bullying Victimization. There are two distinct types of bullying victimization that can be measured: direct and indirect. Direct bullying victimization is measured by four items, including (a) someone threatened to hurt me at school, (b) someone hit me, (c) someone

6

Popp et al.

used strong-arm or forceful methods to get money or things from me, or (d) someone bullied me or picked on me. Indirect bullying is measured by two items, including (a) in class, I often feel “put down” by my teachers; and (b) in class, I often feel “put down” by other students. The measures of bullying victimization were dichotomized to indicate whether the student experienced bullying victimization during the first semester or term of 2001–2002. The reference categories are not being a victim of direct or indirect ­bullying. Gender. Gender is measured as the student’s self-report as male or female. Male was the reference category. Student and School Characteristics. As noted, previous studies have established that both student and school characteristics are associated with bullying victimization, academic self-efficacy, and educational achievement. Student characteristics such as race, ethnicity, and socioeconomic status as well as school characteristics including location, size, and economic affluence are related to bullying victimization, academic self-efficacy, and educational achievement. Students self-reported their race or ethnicity; four racial and ethnic groups were considered: White American (the reference category), Black/African American, Latina/Latino American, and Asian American. Family socioeconomic status is a preconstructed measure that is a standardized (z score) variable based on five equally weighted, standardized components: father’s/guardian’s education, mother’s/guardian’s education, family income, father’s/guardian’s occupational prestige, and mother’s/guardian’s occupational prestige (see NCES, 2004 for further detail). School location is measured by the type of community: urban, suburban (the reference category), or rural. School size is measured as the number of 10th-grade students enrolled in the school. School poverty is measured by the percentage of students who receive free or reduce-priced lunches at the school. Descriptive statistics for academic self-efficacy, educational achievement, and other study variables are reported in Table 1.

Procedures and Analysis of Data Because ELS is designed as a cluster sample in which schools are sampled with unequal probability and then students are sampled within these selected schools, the subsample of ELS would violate the assumption of independent observations. Although hierarchical linear modeling is often used to address the issue of nesting with school survey data, school factors are not central to the research questions. Thus, this study accounts for this nonindependence and nesting by using a survey estimation technique in Stata, which takes into consideration the clustering in the sample design. The survey estimators are adjusted for clustering, stratification, and weighting to ensure a nationally representative sample. Ordinary least squares (OLS) regression analysis is used to examine the relationship between gender, bullying victimization, and educational outcomes (i.e., academic self-efficacy and educational achievement) while controlling for the effects of student and school characteristics. The analyses proceed in several steps. Table 1 presents the descriptive statistics and the results from the analysis of variance (ANOVA). Table 2 displays the results of the regression analysis of the relationships between gender, bullying victimization, and academic self-efficacy. In the baseline model of Table 2, academic self-efficacy is regressed on gender and other student and school characteristics. In the second model, the effect of the bullying victimization measures (i.e., direct and indirect) are analyzed. In the final model, the interactions between gender and bullying victimization are added to the analysis of

Gender, Bullying, and Education7

TABLE 1.  Descriptive Statistics of Education Longitudinal Study of 2002 10th-Grade Sample by Gender Total Variable

M

Male Students SD

M

SD

Female Students M

SD

Educational outcomes   Academic self-efficacy

12.07

8.85

11.73

9.22

12.39

8.45***

  Educational achievement

49.61

10.03

49.39

10.31

49.84

9.74*

  Direct bullying

0.40

0.49

0.46

0.50

0.34

0.47***

  Indirect bullying

0.25

0.43

0.24

0.43

0.26

0.42*

Student bullying victimization

Student and school characteristics   Black/African American

0.16

0.36

0.16

0.36

0.15

0.36

  Latino/Latina American

0.17

0.37

0.17

0.37

0.17

0.37

  Asian American

0.11

0.31

0.11

0.32

0.11

0.32

  White American

0.57

0.49

0.57

0.50

0.57

0.49

  Family SES

–0.07

0.71

–0.07

0.70

–0.09

0.72

  School size

368.80

232.27

371.45

229.55

366.21

234.90

28.67

21.33

28.65

21.11

28.69

21.51

0.27

0.44

0.27

0.44

0.28

0.44

  School poverty  Urban N

10,440

5,120

5,320

Note. SES 5 socioeconomic status. Significant differences between males and females are denoted with asterisks which are based on analysis of variance (ANOVA) tests. *p  .05. ***p  .001. academic self-efficacy. Table 3 presents a similar regression analysis, as just described, but with the relationships between gender, bullying victimization, and educational ­achievement being analyzed.

RESULTS Descriptive statistics for the dependent and independent variables are reported in Table l. The results of the ANOVA analyses indicate that there are statistically significant differences between girls and boys. Girls indicate higher levels of academic self-efficacy (M 5 12.39, SD 5 8.45, p  .001) and educational achievement (M 5 49.84, SD 5 9.74, p  .05) in comparison to boys’ academic self-efficacy (M 5 11.73, SD 5 9.22, p  .001) and educational achievement (M 5 49.39, SD 5 10.31, p  .05). Girls report fewer incidents of direct bullying victimization (M 5 0.34, SD 5 0.47, p  .001) in comparison to boys’ direct bullying victimization (M 5 0.46, SD 5 0.50, p  .001). On the other hand, girls

8

Popp et al.

TABLE 2.  Ordinary Least Squares Regression Results of Selected Predictors on Academic Self-Efficacy (N 5 10,440) Model 1 b

SE

Direct bullying victimization

Model 2 b

SE

20.54**

0.20

 Female Indirect bullying victimization

20.97***

0.22

 Female

Model 3 b

SE

20.29

0.28

20.52**

0.40

20.30

0.32

21.35**

0.44

Student characteristics  Female

0.51**

0.20

0.44*

0.20

0.98***

0.27

  Black/African American

22.06***

0.36

22.12***

0.36

22.08***

0.36

 Latino/Latina

21.73***

0.39

21.72***

0.39

21.72***

0.39

  Asian American

20.61

0.46

20.67

0.46

20.69

0.46

  Family SES

2.24***

0.16

2.20***

0.16

2.19***

0.17

School characteristics  Urban

20.14

0.35

20.15

0.35

20.16

0.35

 Rural

20.37

0.40

20.37

0.40

20.36

0.40

 Size

0.21*

0.09

0.09

20.01

Intercept

12.44

12.95

12.66

0.07

0.09

0.09

20.01

0.01

0.20*

 Poverty R2

0.01

0.20*

20.01

0.09 0.01

Note. SES 5 socioeconomic status. The reference categories are male, no direct or indirect bullying victimization, White American (non-Hispanic), and suburban schools. *p  .05. **p  .01. ***p  .001.

report higher incidents of indirect bullying victimization (M 5 0.26, SD 5 0.42, p  .05) in comparison to boys’ indirect bullying victimization (M 5 0.24, SD 5 0.43, p  .001). These findings are consistent with previous research findings, indicating that there are gender differences related to educational outcomes and bullying victimization. Table 2 displays the results of the linear regression analysis of the relationships between gender, bullying victimization, and academic self-efficacy. Model 1 presents the baseline linear regression analysis. Gender is related to academic self-efficacy; while controlling for other factors, girls report an increase of 0.51 in their academic self-efficacy score compared to boys. Furthermore, family socioeconomic status and school size are positively related to academic self-efficacy; however, being Black/African American or Latino/ Latina American is negatively related to the student’s academic self-efficacy. In Model 2, measures of bullying victimization are included in the analysis. Bullying victimization has a negative link with academic self-efficacy. While controlling for other

Gender, Bullying, and Education9

TABLE 3.  Ordinary Least Squares Regression Results of Selected Predictors on Educational Achievement (N 5 10,440) Model 1 b

SE

Direct bullying victimization

Model 2 b

SE

–1.03***

0.19

 Female Indirect bullying victimization

–2.56***

0.23

 Female

Model 3 b

SE

–0.48*

0.28

–1.14**

0.41

–2.28***

0.31

–0.56

0.48

Student characteristics  Female

0.42*

0.19

  African American/Black

–6.48***

0.36

 Latino/Latina

–4.99***

0.37

0.28

0.19

0.87***

0.25

–6.59***

0.35

–6.55***

0.36

–4.95***

0.37

–4.95***

0.37

  Asian American

0.55

0.49

0.43

0.49

0.41

0.49

  Family SES

4.63***

0.17

4.54***

0.16

4.52***

0.17

School characteristics  Urban

–0.29

0.39

–0.31

0.39

–0.32

0.39

 Rural

0.05

0.35

0.04

0.35

0.05

0.34

 Size

0.05

0.08

0.04

0.08

0.04

0.08

 Poverty

–0.03***

Intercept

52.30

53.46

53.14

0.26

0.28

0.28

R2

0.01

–0.03***

0.01

–0.03***

0.01

Note. SES 5 socioeconomic status. The reference categories are male, no direct or indirect bullying victimization, White American (non-Hispanic), and suburban schools. *p  .05. **p  .01. ***p  .001. predictors of academic self-efficacy, students who report being a victim of direct bullying have a decrease of 0.54 in their academic self-efficacy score in comparison to students who reported not being a victim of direct bullying. Students who report being a victim of indirect bullying have a decrease of 0.97 in their academic self-efficacy score in comparison to students who reported not being a victim of indirect bullying. Being female remains positively related to academic self-efficacy; however, the strength of the relationship is slightly reduced from the previous model. Family socioeconomic status and school size still have a positive association with academic self-efficacy, although being Black/African American or Latino/Latina American continues to have a negative association with academic self-efficacy. In Model 3 of the analysis, the interaction terms for gender and bullying victimization are added. While controlling for other factors, girls who report being a victim of direct bullying have a decrease of 0.52 in their academic self-efficacy score, whereas for boys, being a victim of direct bullying is not associated with their academic self-efficacy. Girls

10

Popp et al.

who report being a victim of indirect bullying have a decrease of 1.35 in their academic self-efficacy score, whereas for boys, being a victim of indirect bullying is not associated with their academic self-efficacy. Not only does being female remain positively associated with academic self-efficacy; while controlling for girls’ bullying victimization, the strength of the relationship increases from the previous model. In other words, girls have an increase of 0.98 reflected in their academic self-efficacy score in comparison to boys. Family socioeconomic status and school size continue to have a positive role with academic self-efficacy, although being Black/African American or Latino/Latina American continues to have a negative effect on academic self-efficacy. Table 3 displays the findings from the regression analysis of gender, bullying victimization, and educational achievement. Model 1 presents the baseline regression model. There is a gender difference in educational achievement. Being female is positively linked with educational achievement; while controlling for other predictors, girls report an increase of 0.42 in their educational achievement in comparison to boys. Family socioeconomic status has a positive effect on educational achievement, and being Black/African American, being Latino/Latina American, and the school poverty have a negative effect on the student’s educational achievement. In Model 2 of Table 3, measures of bullying victimization are now considered in the analysis. Bullying victimization has a negative effect on educational achievement. While controlling for other factors, students who report being a victim of direct bullying have a decrease of 1.03 in their educational achievement score in comparison to students who reported not being a victim of direct bullying. Students who report being a victim of indirect bullying have a decrease of 2.56 in their educational achievement scores in comparison to students who reported not being a victim of direct bullying. Gender does not have a statistically significant effect on educational achievement scores in this model. Family socioeconomic status remains positively related to educational achievement. Being Black/ African American, being Latino/Latina American, and school poverty continue to be negatively linked with educational achievement. In Model 3 of the educational achievement analysis presented in Table 3, the interaction terms for gender and bullying victimization are added. While controlling for other variables, girls who report being a victim of direct bullying experience a decrease of 1.62 reflected in their educational achievement scores, whereas boys who report being a victim of direct bullying have a 0.48 decrease in their educational achievement. The relationship between indirect bullying victimization and educational achievement is not gendered. Both boys and girls who report being a victim of indirect bullying have a decrease of 2.28 reflected in their educational achievement scores. In this model, gender has a statistically significant effect on educational achievement. Girls have an increase of 0.87 reflected in their educational achievement in comparison to boys. Family socioeconomic status has a positive effect on educational achievement, and being Black/African American, being Latino/Latina American, and school poverty have a negative effect on educational ­achievement.

DISCUSSION Gender is consistently associated with educational outcomes; female students have better educational outcomes than male students in terms of both academic self-efficacy and educational achievement. In light of this finding, schools need to consider how to improve

Gender, Bullying, and Education11

boys’ educational experiences so that boys develop a similar level of academic self-­ efficacy and, in turn, educational achievement as their female counterparts. As expected, bullying victimization is negatively associated with educational outcomes. In terms of direct and indirect bullying and academic self-efficacy, there are gender disparities: Girls who are victims of direct and indirect bullying have lower academic selfefficacy, whereas the link between direct and indirect bullying victimization for boys’ academic self-efficacy are not statistically significant. Schools must be more proactive in curtailing both direct and indirect bullying, particularly for girls, because both negatively impact academic self-efficacy. Academic self-efficacy is too important to be undermined by either form of bullying because it is a prerequisite to other educational outcomes (Caprara et al., 2011; Huang, 2013; Young et al., 2006). Bullying victimization negatively impacts the victim’s educational achievement. It appears direct bullying has a gender-specific effect on educational achievement. Interestingly, the detrimental effect of direct bullying victimization appears to be greater for girls than boys. Girls experiencing direct bullying may be more concerned about their personal safety because of the bully’s physical attacks, threats, and/or destruction of the girls’ personal property than about achieving their educational goals. This supports previous research, which suggests that although bullying negatively impacts both genders, girls experience greater psychological distress as a result (Klein, 2012; Young et al., 2006), which in turn is undermining their ability to perform on standardized tests. On the other hand, there is no gender difference in terms of the consequences of indirect bullying victimization on educational achievement. Being a victim of indirect bullying undermines students’ educational achievement for both boys and girls. The verbal attacks, insults, gossiping, and social isolation not only impacts girl but also have a significant impact on boys’ educational achievement. Indirect bullying undermines the students’ psychological well-being, which in turn affects their performance in school. It was unexpected to find that indirect bullying has a negative effect on boy’s educational achievement. Although the study by Dukes et al. (2010) explored the gender distinctions with the effect of direct and indirect bullying on carrying weapons, they also did unexpectedly find that indirect bullying resulted in increased weapon carrying for boys. Dukes et al. suggest that “physical bullying may be a more accepted behavior (especially among adolescent boys), so bonding and social development may be weaker among relational bullies” (p. 527). In other words, because direct bullying victimization is argued to be more common for boys, do boys have the coping skills and/or social support that acknowledges that indirect bullying victimization may be damaging? Hampel, Manhal, and Hayer (2009) indicate that indirect bullying results in maladaptive coping, including rumination for both girls and boys. Gender still impacts educational outcomes in a significant manner, and a better understanding of the mechanism that creates these gender disparities is needed to alleviate these differences. In particular, research should explore how girls are able to develop academic self-efficacy in an environment riddled with hidden curriculum messages and gendered student expectations and yet outperform their male counterparts on many of the measures of educational success (Buchmann et al., 2008; Morris, 2012). In light of the progress made by girls and women in education (Buchmann et al., 2008; Morris, 2012), research needs to explore whether the hidden curriculum is still in place and the extent to which it is a meaningful barrier to the educational success of female students. Because bullying begins prior to high school, more research needs to be conducted in middle schools to determine the consequences of bullying victimization and determine if other demographic disparities in the consequences of bullying exist. Not only are there

12

Popp et al.

gender differences in bullying victimization but there are also racial and ethnic disparities in bullying victimization (Bradshaw et al., 2013; Peguero, 2012; Peguero & Williams, 2011). It would not be surprising if future research were to determine that consequences of bullying are race-specific. It stands to reason that the consequences of intraracial bullying might differ from interracial bullying victimization. Because this analysis did not consider whether the school the student attends has an active antibullying program in place, future research should examine the role of antibullying programs that may indeed ameliorate bullying as well as its detrimental effects. Schools with an antibullying program may mitigate the consequences of bullying on the student and his or her educational outcomes. In addition, future research should explore the role of friendship networks. The research literature on bullying makes it clear that having social support from teachers, administrators, and peers that can come to your aid when you are being bullied is critical to ending bullying victimization (Espelage & Swearer, 2010; Hong & Espelage, 2012); it is reasonable to assume that having friends may be important in determining the educational consequences of bullying. The results of this study indicate that gender and bullying victimization have significant impacts on both academic self-efficacy and educational achievement. In addition, the results suggest that the effect of bullying victimization on educational outcomes is moderated by the student’s gender. Despite the educational gains made by female students, schools are still gendered places, and a student’s gender has consequences in that environment, which needs to be considered further. Last, this study illustrates the importance of differentiating between the forms of bullying and exploring their respective impacts on educational outcomes.

REFERENCES Bandura, A. (1977). Social learning theory. New York, NY: General Learning Press. Bandura, A., Barbaranelli, C., Caprara, G., & Pastorelli, C. (2001). Self-efficacy beliefs as shapers of children’s aspirations and career trajectories. Child Development, 72, 187–206. Bradshaw, C. P., Waasdorp, T. E., Goldweber, A., & Johnson, S. L. (2013). Bullies, gangs, drugs, and school: Understanding the overlap and the role of ethnicity and urbanicity. Journal of Youth and Adolescence, 42, 220–234. Brown, L., & Iyengar, S. (2008). Parenting styles: The impact on student achievement. Marriage & Family Review, 43(1–2), 14–38. Buchmann, C., Condron, D. J., & Roscigno, V. J. (2010). Shadow education, American style: Test preparation, the SAT and college enrollment. Social Forces, 89, 435–462. Buchmann, C., DiPrete, T., & McDaniel, A. (2008). Gender inequalities in education. Annual Review of Sociology, 34(3), 19–37. Caprara, G., Vecchione, M., Alessandri, G., Gerbino, M., & Barbaranelli, C. (2011). The contribution of personality traits and self-efficacy beliefs to academic achievement: A longitudinal study. British Journal of Educational Psychology, 81(1), 78–96. Dukes, R. L., Stein, J. A., & Zane, J. I. (2010). Gender differences in the relative impact of physical and relational bullying on adolescent injury and weapon carrying. Journal of School Psychology, 48, 511–532. Entwisle, D. R., Alexander, K. L., & Olson, L. S. (2007). Early schooling: The handicap of being poor and male. Sociology of Education, 80, 114–138. Espelage, D. L., & Swearer, S. M. (2003). Research on bullying and victimization: What have we learned and where do we go from here? School Psychology Review, 32, 365–383. Espelage, D. L., & Swearer, S. M. (2010). Bullying in North American schools. New York, NY: Routledge Press.

Gender, Bullying, and Education13 Faris, R., & Felmlee, D. (2011). Status struggles network centrality and gender segregation in sameand cross-gender aggression. American Sociological Review, 76, 48–73. Finkelhor, D. (2008). Childhood victimization: Violence, crime and abuse in the lives of young people. New York, NY: Oxford University Press. Hampel, P., Manhal, S., & Hayer, T. (2009). Direct and relational bullying among children and adolescents: Coping and psychological adjustment. School Psychology International, 30, 474–490. Hanson, T., Austin, G., & Zheng, H. (2010). Academic performance and school well-being in California. California Healthy Students Research Project Research Brief. No. 1. San Francisco, CA: WestEd. Hong, J. S., & Espelage, D. L. (2012). A review of research on bullying and peer victimization in school: An ecological systems analysis. Aggression and Violent Behavior, 17, 311–322. Huang, C. (2013). Gender differences in academic self-efficacy: A meta-analysis. European Journal of Psychology of Education, 28, 1–35. Juvonen, J. (2011). Bullying and violence as barriers to academic achievement. California Healthy Students Research Project Research Brief. No. 4. San Francisco, CA: WestEd. Juvonen, J., Wang, Y., & Espinoza, G. (2011). Bullying experiences and compromised academic performance across middle school grades. Journal of Early Adolescence, 31, 152–173. Kao, G., & Thompson, J. S. (2003). Racial and ethnic stratification in educational achievement and attainment. Annual Review of Sociology, 29, 417–443. Kingston, P., Hubbard, R., Lapp, B., Schroeder, P., & Wilson, J. (2003). Why education matters. Sociology of Education, 76, 53–70. Klein, J. (2012). The bully society: School shootings and the crisis of bullying in America’s schools. New York, NY: New York University Press. Malone, L., West, J., Denton, K., & Park, J. (2006). The early reading and mathematics ­achievement of children who repeated kindergarten or who began school a year late. Washington, DC: National Center of Educational Statistics. Morris, E. W. (2012). Learning the hard way: Masculinity, place, and the gender gap in education. News Brunswick, NY: Rutgers University Press. Nakamoto, J., & Schwartz, D. (2010). Is peer victimization associated with academic achievement? A meta-analytic review. Social Development, 19, 221–242. National Center for Education Statistics. (2004). Educational longitudinal study: 2002 base year data file user’s manual. Washington, DC: Author. Olweus, D. (1993). Bullying at school: What we know and what we can do. Cambridge, MA: Blackwell. Pajares, F. (2008). Motivational role of self-efficacy beliefs in self-regulated learning. In D. H. Schunk & B. J. Zimmerman (Eds.), Motivation and self-regulated learning: Theory, research, and applications. New York, NY: Lawrence Erlbaum Associates. Peguero, A. A. (2012). Schools, bullying, and inequality: Intersecting factors and complexities with the stratification of youth victimization at school. Sociology Compass, 6(5), 402–412. Peguero, A. A., & Williams, L. A. (2011). Racial and ethnic stereotypes and bullying victimization. Youth & Society. Advance online publication. Popp, A. M., & Peguero, A. A. (2011). Routine activities and victimization at school: The significance of gender. Journal of Interpersonal Violence, 26(12), 2413–2436. Schwartz, D., Gorman, A. H., Nakamoto, J., & Toblin, R. L. (2005). Victimization in the peer group and children’s academic functioning. Journal of Educational Psychology, 97, 425–435. Swearer, S. M. (2008). Relational aggression: Not just a female issue. Journal of School Psychology, 46, 611–616. Tach, L., & Farkas, G. (2006). Learning-related behaviors, cognitive skills, and ability grouping when schooling begins. Social Science Research, 35(4), 1048–1079. Urdan, T., & Pajares, F. (2006). Self-efficacy beliefs of adolescents. Charlotte, NC: Information Age. Wei, H., & Williams, J. H. (2004). Relationship between peer victimization and school adjustment in sixth grade students: Investigating mediation effects. Violence and Victims, 19, 557–571.

14

Popp et al.

Wilcox, P., Tillyer, M. S., & Fisher, B. S. (2009). Gendered opportunity? School-based adolescent victimization. Journal of Research in Crime and Delinquency, 46, 245–269. Wood, D., Kaplan, R., & McLoyd, V. (2007). Gender differences in the educational expectations of urban, low-income African American youth: The role of parents and the school. Journal of Youth & Adolescence, 36(4), 417–427. Young, E., Boye, A., & Nelson, D. (2006). Relational aggression: Understanding, identifying, and responding in schools. Psychology in the Schools, 43(3), 297–312. Correspondence regarding this article should be directed to Anthony A. Peguero, PhD, Department of Sociology, Virginia Tech, Blacksburg, VA 24061. E-mail: [email protected]

Violence and Victims, Volume 16, Number 4, 2001

Short-Term Stability and Prospective Correlates of Bullying in MiddleSchool Students: An Examination of Potential Demographic, Psychosocial, and Environmental Influences Dorothy L. Espelage University of Illinois at Urbana-Champaign

Kris Bosworth The University of Arizona

Thomas R. Simon Centers for Disease Control and Prevention Stability and change of bullying over a four-month interval was examined in 516 middle school students (grades 6–8). The stability coefficient was .65 for the entire sample. There was a significant increase in bullying behavior from Time 1 to Time 2 for 6th grade students; no significant change in bullying was found among 7th or 8th graders. For 6th graders, a greater confidence in using non-violent strategies was associated with less bullying at Time 2, while beliefs supportive of violence and misconduct, less positive adult influences, and more negative peer influences were associated with greater likelihood of bullying at Time 2. Higher levels of impulsivity, anger, and depression were also associated with greater levels of bullying over time. Several explanations for the increase in bullying behaviors among 6th graders are discussed and linked to intervention efforts.

R

ecent violent events in U.S. schools highlighted by the news media have involved students who were allegedly victims of bullying (Begley, 1999). As such, parents and educators have become more concerned about these low-level aggressive behaviors during childhood and adolescence that were previously seen as normative and harmless. While a plethora of research has been conducted on physical, overt aggression within U.S. schools, fewer studies have focused specifically on low-level aggression such as the teasing, name-calling, and threatening that characterize bullying. For the purposes of this study, bullying is defined as a subset of aggressive behavior that has potential to cause physical or psychological harm to the recipient, and includes name-calling, teasing, verbal threats, social exclusion, and pushing (Hoover, Oliver, & Hazier, 1992; Rigby, Cox, & Black, 1997; Thompson & Sharp, 1998). Although many studies on bullying have been conducted outside the United States, several recent reports provide some insight into the prevalence of bullying within U.S.

© 2001 Springer Publishing Company15

16

D. L. Espelage et al.

schools. Available data from cross-sectional studies in midwestern and southeastern U.S. schools suggest that bullying behavior is quite common. In a study of junior high and high school students from midwestern towns, 88% reported having observed bullying and 77% reported being a victim of bullying during their school years (Hoover et al., 1992). Similarly, 25% of students in grades 4 through 6 admitted to bullying another student with some regularity in the three months preceding the study (Limber et al., 1997). A primary aim of this study is to examine whether bullying behaviors remain stable over a 4-month time period among middle-school students (grades 6–8). During early adolescence, the function and importance of the peer group changes dramatically (Crockett, Losoff, & Petersen, 1984; Dornbusch, 1989). Adolescents turn to their peers to discuss problems, feelings, and fears, thereby the time spent with friends becomes very important (Sebald, 1992; Youniss & Smollar, 1985). With this increased reliance on peers for social support comes pressure to attain social status (Corsaro & Eder, 1990; Eder, 1985). The pressure to obtain peer acceptance and status is likely to be intense as young adolescents attempt to negotiate a complex social environment that presents unique challenges not encountered during the elementary school years. Recent studies have actually found that bullying among 6th graders might be associated with popularity and for some students it serves to enhance their social standing within classrooms (Espelage & Holt, in press; Pellegrini, Bartini, & Brooks, 1999). Therefore, we hypothesized that bullying behaviors might escalate over the course of a semester for middle-school students. Our examination of bullying differs in several ways from previous investigations of aggression during early adolescence. Given the results of ethnographic analyses of the middle-school culture, which found that teasing is more frequent than overt aggression (Eder, 1995), we deviate from the mainstream aggression literature by including verbal teasing and threatening behaviors in addition to overt, physical aggression. Second, our goal is not to identify students who “bullied the most,” but to assess bullying as a continuum of behavior from low to high levels. Missing from studies that categorize students as “bullies” or “victims” is a recognition that many students might tease their peers in more subtle ways and on a less regular basis and that such behavior can be harmful (Bosworth, Espelage, & Simon, 1999). Finally, unlike previous studies on bullying behavior where students were presented with a definition of “bullying” (e.g., Boulton & Underwood, 1992; Salmivalli, Lagerspetz, Bjorkqvist, Osterman, & Kaukiainen, 1996), we ask students about the frequency of behaviors (e.g., teasing, making fun of others) without requiring them to evaluate the impact or motivation of their behavior. We feel that this approach minimizes socially desirable responding. This is particularly important because students who tease others often do not define these behaviors as bullying or even consider them hurtful, despite the findings that victims report significant distress from low-level verbal aggression (Hoover & Oliver, 1996; Hoover, Oliver, & Thompson, 1993).

Correlates of Stability and Change in Bullying In addition to documenting the stability and change in bullying behaviors across middleschool students, this study examines the demographic, psychosocial, familial, and environmental correlates associated with bullying behavior over time. The commonly cited developmental model of childhood aggression proposed by Patterson and colleagues (1992) suggests that certain familial factors can interact with individual characteristics of a child to lay the foundation for aggressive interactions with peers and teachers and place the child

Stability of Bullying in Middle School17

at risk for gravitating toward deviant peers. Thus, to represent the complex and multidetermined nature of bullying, we include a wide range of variables. Drawing from the extant literature on aggression and the developmental model of aggression (Patterson, Reid, & Dishion, 1992), study variables are organized into three categories, including demographic variables, psychosocial characteristics, and familial and environmental factors. In our previously published work examining correlates of bullying behavior at Time 1 (Bosworth et al., 1999; Espelage, Bosworth, & Simon, 2000), these variables have been significantly associated with bullying behavior, and therefore are retained in this longitudinal investigation. Demographic Variables. Previous research has yielded inconsistent findings on the differences between males and females in the stability and change of aggression. Cairns and colleagues (1989) found no sex differences in the stability of teacher-rated and self-rated aggression from 4th to 8th grade; but they found lower stability in girls’ aggression at the 9th grade assessment. Similarly, in a 22-year longitudinal study, Huesmann and colleagues (1984) reported stability in aggression for males and females from age 8 through early adulthood; although the differences between those high on aggression versus those low or medium on aggression were less pronounced for females than they were for males. Given these inconsistent findings, we examine the impact of sex and grade on the stability and change in bullying behaviors. In addition, the impact of other demographic variables is explored. Psychosocial Characteristics. Factors that have been linked to aggression during early adolescence were selected for inclusion in the study, but a lengthy description of the literature implicating these factors would be cumbersome here. Thus, we refer the reader to our previously published work for a more detailed discussion (Bosworth et al., 1999; Espelage et al., 2000). In the current study, variables consistently associated with bullying and aggression are evaluated as predictors of bullying over time, including misconduct (Klicpera & Klicpera, 1996), anger (Dodge, 1991), depression (Dumas, Neese, Prinz, & Blechman, 1996; Slee, 1995; Slee & Rigby, 1993), impulsivity (Olweus, 1994; Pope & Bierman, 1999), and sense of belonging at school (Klicpera & Klicpera, 1996). Furthermore, to elucidate the role of violence-related attitudes and beliefs in the stability and change of bullying behavior, we also include variables which are often targeted in violence prevention programs, including beliefs supportive of violence (Bentley & Li, 1995), confidence in using non-violent strategies, and intentions to use non-violent strategies (Bosworth et al., 1999). Familial and Environmental Variables. Development of bullying and aggression often involves influences from the larger social environment, including modeling of aggressive behaviors by adults and deviant behaviors of the child’s peers (Gorman-Smith, Tolan, Zelli, & Huesmann, 1996; Thomberry, 1994; Tolan, Cromwell, & Braswell, 1986). Therefore, we explore the impact of familial and peer influences on stability and change of bullying behavior by assessing the messages students receive from their parents about violence; examining parents’ physical discipline strategies; and assessing the amount of time each student spends with his or her family and the extent to which he or she is supervised (Farrington, 1991). Finally, safety concerns at school and in the neighborhood and access to guns are evaluated as potential correlates (Gorman-Smith & Tolan, 1998; Webster, Gainer, & Champion, 1993).

Study Questions The study addresses several questions. First, we question whether bullying behavior remains stable over a four-month time interval or whether it increases or decreases. Second, we assess whether change in bullying behavior differs by demographic factors such as sex,

18

D. L. Espelage et al.

grade, ethnicity, family type, and lunch status. Finally, we seek to evaluate the extent to which individual characteristics, psychosocial influences, familial and environmental factors, and bullying behaviors at Time 1 are associated with bullying at the 4-month follow-up.

METHOD Participants The data for this study were collected as part of a larger evaluation of a multimedia violence prevention study (Bosworth, Espelage, Dubay, Dahlberg, & Daytner, 1996). Students from one large middle-school completed a survey in January 1995 (Time 1) and completed a similar survey in May 1995 (Time 2). A detailed description of the evaluation design, survey administration, and preliminary analyses can be found in Bosworth and colleagues (1996). While the multimedia prevention program had a small significant effect in positively changing potential mediating factors associated with aggression (e.g., beliefs supportive of violence), it failed to impact levels of aggressive behavior (Bosworth, Espelage, Daytner, DuBay, & Karageorge, 2000). However, to control for the influence of the prevention program, intervention status (i.e., intervention versus control) was entered as a covariate in the regression analyses. At Time 1, parent permission forms were sent to all 1,361 students registered at this school. Of the 700 (51%) who returned permission forms, 142 denied permission, leaving a sample of 558. To determine the representativeness of our study sample at Time 1, chi-square analyses were conducted to test differences in relevant demographic variables between the 558 study participants and the students who did not return a permission form (nonparticipants), including sex, grade, ethnicity, family type, and free/reduced-fee lunch. A significant difference was found for the distribution of grades between the two groups; approximately 42% of the study participants were 6th graders, while 25% of the nonparticipants were 6th graders. No other significant differences were found between study participants and non-participants (Bosworth et al., 1996). Of the 558 students who completed the survey at Time 1, 538 also completed it at Time 2. Three students refused to take the postsurvey; thirteen students had moved; three had been expelled from school; and one had a severe medical condition. Data from 22 students were excluded because of significant missing data or random answering patterns. Thus, the panel comprised 516 students. Of the 516 students in the panel, 54% were females and 46% were males, with 42% 6th graders, 30% 7th graders, and 28% 8th graders. Approximately 84% were Caucasian, 9% were African American, 3% were biracial, and 4% reported other racial backgrounds. Fifty percent reported living with two biological or adoptive parents (two parents, no step parents); 20% reported living with two parents, one of whom was a step-parent; 28% reported residing with a single parent; and the remaining 2% reported other living arrangements (e.g., grand-parents, foster care). In addition, 28% of the sample were receiving free or reduced-priced lunch, and 17% had chapter 1 status indicating qualification for remedial academic support.

Measures Existing measures with strong psychometric properties were selected from a comprehensive literature review and adapted for this study. All measures were presented to groups of

Stability of Bullying in Middle School19

middle-school students for their review. Based on the results from these student groups, items were modified for clarity and readability. Exploratory factor analysis was then conducted for all study measures. Factors were extracted based on eigen values, percent of variance explained, and examination of scree plots. Items that had factor loadings above .50 and items that did not have cross-loadings above .30 on any other factor were retained. We refer the reader to sources in which the psychometric properties of each measure are described in greater detail (Bosworth et al., 1999; Dahlberg, Toal, & Behrens, 1998; Espelage et al., 2000).

Demographic Characteristics The sociodemographic characteristics of sex, grade, ethnicity, and free/reduced-fee lunch were assessed using self-reports. In addition, students were asked “Who lives with you?” Based on these responses, students were categorized into four mutually exclusive caregiver types: two biological/adoptive parents, step-family, single parent and other (e.g., grandparents, foster care).

Bullying Behaviors Bullying behavior was assessed with a scale that consisted of five items from the Modified Aggression Scale (Dahlberg et al., 1998). The items on this scale were consistent with the definition of bullying used in the current study and reflected both psychological and physical aspects of bullying. Participants were asked how many times they did the following in the last 30 days: (a) “I called other students names”; (b) “I teased students”; (c) “I said things about students to make other students laugh”; (d) “I threatened to hit or hurt another student”; and (e) “I pushed, shoved, slapped, or kicked other students.” Response choices included 0 = Never, 1 = 1 or 2, 2 = 3 or 4, and 3 = 5 or more times (Time 1  = .86; Time 2  = .86). Summing across all five items created total scores ranging from 0 through 15, with 0 meaning no bullying activity.

Psychosocial Characteristics Misconduct. Misconduct was measured with a six-item scale developed specifically for this study. The first three questions presented to participants required them to indicate how many times in the last 30 days they broke a rule:

1. At school?; 2. At home?; and 3. In the community?

The last three questions required participants to indicate how many times in the last 30 days they broke the law in these same locations. Response choices ranged from 0 = Never through 4 = 7 or more times ( = .79). A log transformation of the scale score was calculated and used in subsequent analyses because the distribution was highly skewed with the majority of students reporting few incidences of breaking a rule or law. Anger. The Anger Scale consisted of one item from the University of Texas-Houston Health Science Center Student Questionnaire (Dahlberg et al., 1998), and three items developed specifically for our investigation. First, respondents were asked two questions about their anger in the last 30 days (i.e., “I was angry most of the day”; “I took my anger out on an innocent person”). Response choices ranged from 0 = Never through 3 = 5 or

20

D. L. Espelage et al.

more times. In addition, student participants were asked how often they agreed with two additional statements (i.e., “I frequently got angry”; “I was grouchy or irritable, or in a bad mood, so that even little things would make me mad.”). Response options for these items ranged from 0 = Never through 4 = Always. Because the response options were dissimilar across the four items, individual item scores were converted to z-scores. These standardized variables were then used to calculate a Cronbach’s alpha for the scale ( = .70) and z-scores were also summed so that higher scores indicated more anger. Depression and Impulsivity. Self-reported feelings of depression were assessed with a five-item scale from the University of Texas Depression Scale (Dahlberg et al., 1998). For example, students were asked how often in the last 30 days: (a) “Were you very sad?”; and (b) “Did you feel like not eating or eating more than usual?” A four-item scale developed specifically for this investigation was used to assess impulsivity. For example, students were asked how often they would say the following statements about themselves: (a) “I do things without thinking” and (b) “I need to use a lot of self-control to stay out of trouble.” Response choices for both the depression and impulsivity scales ranged from 0 = Never through 4 = Always (Depression  = .74; Impulsivity  = .62). School Sense of Belonging. Perceived belonging at school was assessed with four items from the Psychological Sense of School Membership Scale (Goodenow, 1993). Students were asked how much they agreed with statements such as “I feel proud of belonging to my school.” Response choices ranged from 0 = Strongly Disagree through 4 = Strongly Agree ( = .63). Confidence. A five-item Confidence Scale was developed specifically for the violence prevention investigation from which these data were drawn. To assess confidence in using nonviolent strategies, each student was asked how confident he or she was about doing such things as staying out of fights and talking out a disagreement. Response choices were 0 = Not at all Confident through 4 = Very Confident ( = .85). Intentions to Use Non-Violent Strategies. An eight-item Intentions to use the Nonviolent Strategies Scale was developed for this investigation to measure the respondent’s intentions to use non-violent strategies in a future anger-provoking situation. Students were asked how likely they would use strategies such as ignoring the situation and channeling anger into something constructive. Response choices included 0 = Very Unlikely through 3 = Very Likely ( =.63). Beliefs Supportive of Violence. The Beliefs Supportive of Violence Scale was a modified version of a scale from the Houston Community Project Scale (Dahlberg et al., 1998). To assess participants’ beliefs supportive of violence, participants were asked how much they agreed or disagreed with each of six statements, such as “If I walked away from a fight, I’d be a coward.” Response choices ranged from 0 = Strongly Disagree through 4 = Strongly Agree ( = .71).

Familial and Environmental Influences Positive Adult Messages About Violence. Four items selected from the University of Texas-Houston Health Science Center Student Questionnaire (Dahlberg et al., 1998) were used to assess students’ report of what adults tell them about fighting. Respondents were asked to think about the adults with whom they spend the most time and indicate how many of them tell them things like “If another student hits you, hit them back.” Response choices ranged from 0 = None through 3 = All ( = .77). Family Physical Discipline. Students were asked, “If you break a rule in your home, how often are you spanked, hit, or slapped?” Response choices included 0 = Never through

Stability of Bullying in Middle School21

4 = Always. The majority of students indicated that they are never spanked, hit or slapped when they break a rule at home. Thus, a dichotomous variable was created representing two groups:

1. Those students who reported that they never or seldom are spanked, hit or slapped when they break a rule in their home, and 2. Those students who reported that they sometimes, often, or always are spanked, hit, or slapped when they break a rule in their home.

Adult Contact and Time With Family. Time spent without adult supervision was assessed with one item, “On an average weekday, how many hours a day do you spend without an adult around you?” In addition, students responded to the following item to determine how much time they spent with family members, “On an average weekday, how many hours a day do you talk to or do activities with your family?” For both items, response options included 0 = 0 minutes, 1 = 1 to 30 minutes, 2 = 30 minutes to 1 hour, 3 = 1 to 2 hours, and 4 = 2 to 4 hours. Negative Peer Influences.Four items were presented to the study participants. They were asked, “Over the last 30 days, how many of the friends you spent most of your time with did the following?” Example items included “suggested that you do something against the law” and “hit or threatened to hit someone.” Response choices ranged from 0 = None through 3 = All ( = .79). Neighborhood Safety.Participants were asked how often they would agree with the following statements: “I see gang activity in my neighborhood” and “I live in a safe neighborhood.” Response choices included 0 = Never through 4 = Always. A Pearson correlation of -.46 (p < .001) was found between these two items. Response options were recoded for the second item, such that higher scores indicated that the participant believed his/her neighborhood to be unsafe. Both items were then summed; higher scores reflected greater neighborhood safety concerns. Access to Guns.Students were asked how much they agreed or disagreed with the statement “I can get a gun easily.” Response choices ranged from 0 = Strongly Disagree through 4 = Strongly Agree. The majority of students strongly disagreed with this statement. Thus, a dichotomous variable was created representing two groups:

1. Those students whose responses were Neutral, Disagree, or Strongly Disagree, and 2. Those students whose responses were Agree or Strongly Agree.

Feeling Unsafe at School.School safety was measured with a composite score of two items. Students were asked how often they agreed with the following statements: “I worry about my safety getting to and from school” and “I worry about my safety in school.” Five response choices offered ranged from 0 = Never through 4 = Always. A Pearson correlation of .51 (p < .001) was found between these two items.

Data Analysis To examine the short-term stability of bullying, bivariate correlations were calculated between the continuous measures of bullying at Time 1 and Time 2 for the entire sample, and then for each sex and grade. This was followed with four separate three-way mixed ANOVAs to determine whether the change in bullying from Time 1 to Time 2 varied as a function of sex, grade, ethnicity, family type, and lunch status. In each ANOVA, Time was the repeated factor and sex was a between-subjects factor. The second betweensubjects factor was either grade (6th, 7th, or 8th), ethnicity (Caucasian, Non-Caucasian),

22

D. L. Espelage et al.

family type (intact, other), or lunch status (reduced/free fee, no reduction). Next, a series of regression analyses were calculated to examine the extent to which individual characteristics, psychosocial influences, and familial and environmental factors at Time 1 were predictive of bullying behavior at Time 2. Each Time 1 measure was tested in a separate model that controlled for bullying behavior at Time 1 and controlled for whether the participant completed the intervention program or whether the participant was in the control group.

RESULTS Stability of Bullying Behavior The distributions of the Bullying Scale at Time 1 and Time 2 in the current sample were skewed, with the highest percentage of students reporting low levels of bullying behavior (Figure 1). Therefore, these raw scores were subjected to a log transformation, and used as bullying sum scores in subsequent analyses. These transformations resulted in significant decreases in the skewness of the distributions of bullying scores at Time 1 and Time 2. Next, the short-term stability of self-reported bullying behavior was evaluated with correlation coefficients. Transformed bullying scale scores at Time 1 and Time 2 were correlated for the entire sample, and correlations were then calculated separately for each sex and grade. The stability coefficient was .65 (p < .001) for the entire sample. For male and female participants, the associations were r = .65 and r = .64, ps < .001, respectively. Stability coefficients for 6th, 7th, and 8th graders were r = .60, r = .65, r = .73, ps < .001, respectively. Results from the ANOVA models revealed a significant main effect for time, F (1, 512) = 10.39, p < .001, indicating a change in bullying behavior from Time 1 to Time 2 for the entire sample. Significant main effects were also found for sex, F (1, 512) = 14.56, p < .001, and grade, F (2,512) = 3.34, p < .05, collapsing across time points. However, the interaction effect of time and grade (indicating that change in bullying behavior over time varied by grade level) was significant, F (2, 512) = 4.48, p < .05 (Figure 2). To further investigate grade differences in the shift found in bullying behavior, paired post hoc t-tests were carried out separately for each of the three grade levels. As can be seen in Figure 2, 6th graders had a significant increase in bullying behavior from Time 1 (M = 4.16, SD = 3.89) to Time 2 (M = 5.27, SD = 4.32), t (213) = 4.52, p< .001. No significant change in bullying was found among 7th graders from Time 1 (M = 5.44, SD = 4.55) to Time 2 (M = 5.75, SD = 4.39), t (155) = 1.22, p > .05, and no significant change was found for 8th graders (Time 1: M = 4.55, SD = 4.26; Time 2: M = 4.44, SD = 4.06), t (145) = .23, p > .05). The interaction between time and sex was not significant, indicating that the change in bullying was similar for males and females. A three-way interaction for time, grade, and sex was also not significant, which suggests that the interaction between bullying and grade is consistent for males and females (ps > .05). None of the two-way interaction effects between time and ethnicity, family type, and lunch status was statistically significant, Fs (1, 512) = 1.39, 3.58, 2.60, respectively, and the three-way interaction effects among time, sex, and these three demographic variables were also not significant, Fs (1, 512) = .00, 2.26, .28, respectively. Thus, change in bullying behavior over time did not vary as a function of ethnicity, family type, or lunch status or as a function of the interactions of sex with these variables.

Stability of Bullying in Middle School23

Prospective Correlates of Bullying Behavior Among 6th Graders Next, we investigated whether factors in addition to bullying at Time 1 were associated with subsequent reports of bullying behaviors. These analyses included the 6th graders only for several reasons. First, we focused on 6th graders because they were the only participants to report a significant change in bullying from Time 1 to Time 2. Second, the high correlation between bullying at Time 1 and Time 2 for 7th and 8th graders would minimize the impact of any additional variables significantly predicting bullying over time. Third, 6th grade appears to be a crucial time in which bullying might increase as a result of peer group demands. In these analyses, we evaluated the association between demographic characteristics and bullying at Time 2, after controlling for bullying at Time 1 and intervention status. Demographic variables were entered simultaneously into the regression model as predictors, along with Time 1 transformed bullying scale scores and the intervention variable (intervention versus control). Several demographic variables were recoded because of small sample sizes for some categories. Ethnicity was collapsed into two categories (Caucasian, Non-Caucasian), as were caregiver type (intact, other) and lunch status (reduced/free-fee, no reduction). Regression analyses indicated that bullying at Time 1 was significantly associated with bullying at Time 2 for 6th graders ( = .62, p < .001), but intervention status ( = .05), sex ( = .03), ethnicity ( = -.03), family type ( = -.01), and lunch status ( = -.10) were not significantly associated with bullying at Time 2 (ps > .05).

20

time 1

time 2

18

Percentage of Respondents

16 14 12 10 8 6 4 2 0

0

1

2

3

4

5

6 7 8 9 Bully Scale Score

10

Figure 1.  Percentage of bullying scale scores at Time 1 and Time 2.

11

12

13

14

15

24

D. L. Espelage et al. 12

6th

7th

8th

10

Mean Scale Score

8

6

4

2

0 Time 1

Time 2 Time Period

Figure 2.  Bullying scale scores at Time 1 and Time 2 across grades 6–8.

Next, in a series of regressions we examined the influence of psychosocial characteristics, environmental factors, and other contextual variables on 6th graders’ bullying behavior at Time 2, after controlling for bullying behavior at Time 1 and intervention status. The first regression included these two variables as predictors of bullying behavior at Time 2. These variables accounted for 36% of the variance, with bullying at Time 1 significantly related to bullying at Time 2 ( = .60, p < .001) and intervention status was not significantly related to bullying at Time 2 ( = .05, p > .05). In the subsequent regression models, several individual characteristics at Time 1 were associated with bullying behavior at Time 2 after taking into account bullying behavior at Time 1 and intervention status and each accounted for additional variance in the model (see Table 1). First, the strongest relationship was found between confidence to use nonviolent strategies in future conflicts ( = -.25), with a greater confidence in using non-violent strategies associated with less bullying at Time 2. In addition, misconduct ( = .18) and beliefs supportive of violence ( =. 17) were related to bullying behavior at Time 2. These results indicated that more acts of misconduct and greater beliefs supportive of violence at Time 1 were positively associated with the likelihood of bullying at Time 2, after controlling for bullying level at Time 1 and intervention status. Individual characteristics of impulsivity ( = .14), anger ( = .13), and depression ( = .12) were also positively associated with level of bullying over time. Only two environmental characteristics from Time 1, positive adult influence ( = -.15) and negative peer influence ( = .17), were predictive of bullying behavior at Time 2. Less positive adult influences and more negative peer influences at Time 1 were associated with greater bullying at Time 2. Intervention status was not a significant predictor in any of the regression models.

Stability of Bullying in Middle School25

DISCUSSION Over 500 middle-school students from one school during a 4-month period were participants in this study. Their self-reported bullying behavior was measured at two time points. For most students, bullying behavior remained constant between January and May testing times, a finding consistent with previous research in this area (Boulton & Smith, 1994; Dumas et al., 1996; Tomada & Schneider, 1997). An exception was the 6th-grade cohort, where an increase in bullying behavior nearly to the level of the 7th graders was identified. One might speculate that the increase in bullying among 6th graders was a result of their assimilating into the culture of this middle-school, which already had a certain rate of bullying behavior, as was seen in the initial scores for 7th and 8th graders. This is supported by the contention that bullying is a learned behavior and that the 6th graders entering middle-school have not yet learned how to interact in the social milieu of middleschool. Many 6th graders who wish to “fit in” might learn the culture and the behaviors of those who have been in this school longer and have more of the power to dictate the culture (Pellegrini et al., 1999). Although 6th graders in this school were physically isolated from the upper-classmen, they did have contact with upper-classmen on bus rides (usually 30 minutes to an hour) or during assemblies and other planned activities. These times might have provided 6th graders with opportunities to observe teasing and other harassing behaviors. Given recent investigations in which aggressive boys were rated as some of the most popular students by their peers (Espelage & Holt, in press; Rodkin, Farmer, Pearl, & VanArcher, 2000), it is also likely that the 6th graders in our study might have witnessed older children being reinforced for their behavior, thereby increasing the likelihood of these behaviors being imitated. An increase in bullying among 6th graders could be attributed to these young adolescents’ adopting strategies to obtain power and status within their own grade cohort. The transition from elementary school to middle-school can cause stress that promotes this type of behavior as students attempt to define their place in the new social structure. For example, changing from one school to another often leads to an increase in risk-taking behaviors, which might be another way that young people deal with the stress of a new environment, new people, and new rules. The fact that misconduct in school and in the community was associated with bullying behavior in 6th graders suggests that bullying might in fact be part of a spectrum of problem behaviors that emerge in an attempt to negotiate the middle-school social environment. Future research will be necessary to determine the role that this transition plays in the increase in bullying behavior and other disruptive behaviors. Increases in teasing and other harassing behavior is certainly a cause for concern. However, this study was conducted to identify those factors related to bullying over time that could be included in school-based prevention and intervention programs. While bullying at the first time point was most predictive of bullying at the second time point, several other factors emerged as significant correlates of bullying. Consistent with previous analyses of these data within time points (Bosworth et al., 1999), being angry was a significant predictor of bullying over the study period. This replication in a prospective design supports the contention that prevention and intervention programs designed to decrease bullying should include training in anger management. In addition, depressive symptoms and impulsivity also emerged as significant prospective correlates of bullying in 6th graders, suggesting that mood and impulsive behaviors might contribute to the risk of bullying others, which has been found in the literature (Slee, 1995; Slee & Rigby, 1993).

26

D. L. Espelage et al.

Admittedly, structured clinical interviews were not conducted, so these psychological characteristics cannot be assumed to represent a psychiatric diagnosis or even serious mental health symptoms. Nevertheless, bullying might be a type of behavioral coping strategy to manage the emotions of anger, impulsivity, and depression. Regardless of the exact mechanism underlying these associations, the fact that these relations remained after controlling for bullying at baseline suggests that they should be considered as important factors when designing school-based bullying prevention programs. On a positive note, students who had more confidence in their ability to use non-violent strategies and a negative attitude toward using violence as a problem-solving strategy were less likely to report bullying others (Bosworth et al., 1999). The association between bullying and confidence suggests the potential benefit of a direct intervention approach in which students are taught non-violent strategies to solve conflicts and provided ample opportunities (e.g., role-playing) to build confidence in implementing these new learned behaviors.

TABLE 1.  Separate Regression Analysis for 6th Graders Where Individual Characteristics, Environmental Factors, and Other Contextual Variables Predict Bullying Behavior at Time 2 (N = 214)a Predictor Variable

t

R2b

b

SE(b)



Misconduct

 .25

.09

 .18

2.94**

.40

Anger

 .02

.01

 .13

2.11*

.38

Depression

 .01

.01

 .12

2.34*

.38

Impulsivity

 .01

.01

 .14

2.29*

.38

School sense of belonging

-.01

.01

-.08

-1.47

.37

Confidence

-.02

.01

-.25

-3.88***

.41

Intentions

-.01

.01.

-.10

-1.74

.37

Beliefs supportive of violence

 .01

.01

 .17

2.73**

.39

Positive adult models

-.02

.01

-.15

-2.60**

.39

Physical discipline

-.06

.05

-.07

-1.26

.37

Time without adult

 .01

.02

 .03

.59

.37

Time with family

-.01

.02

-.04

-.64

.37

Negative peer influence

 .03

.01

 .17

2.89**

.39

Access to guns

 .09

.05

 .09

1.70

.37

Neighborhood safety

-.01

.01

-.07

-1.25

.37

School safety

-.01

.01

-.02

-.35

.37

aMean Bullying Scale scores at Time 1 and Intervention Status (Intervention versus Control) were entered as covariates in each regression analysis. bRegression model with bullying scale scores at Time 1 and Intervention Status predicting Time 2 Bullying Scale scores yielded an R2 = .36. *p < .05. **p < .01. ***p < .001.

Stability of Bullying in Middle School27

In contrast, intentions to use nonviolent strategies were not significantly associated with reduced bullying over time once previous bullying levels were considered. However, this slight association was in the right direction, suggesting that intentions should not be eliminated as an important predictor of bullying and, therefore, a potential element of bullying prevention efforts. Environmental influences were also considered as prospective correlates of bullying behaviors in 6th graders. As with other risk-taking behaviors, associating with peers with conduct problems is associated with higher levels of bullying behavior. Recalling that misconduct and bullying co-occur in this sample, bullying could be considered part of a spectrum of problem behavior. As such, it is not surprising that a student who bullies his/ her peers has friends who are engaged in similar antisocial activities. Given the suggestion of Salmivalli and colleagues (1996) that bullying is a group process rather than a simple dyadic interaction, an important next step is to determine the extent to which peer group members—either as observers or reinforcers—contribute to bullying behavior. As for adult influences, the amount of time spent with adults appeared to play less of a role in youths’ bullying than did the actual messages about solving conflicts students received from adults. This finding suggests that parents and adults who promote nonviolent strategies do have a positive impact on adolescents’ bullying behaviors. Therefore, parents who promote an open dialogue with their children and adults who have ongoing interactions with students in schools might be critical factors in reducing bullying behavior. Several other environmental characteristics (e.g., neighborhood and school safety, access to guns, parental discipline) were not significantly associated with bullying over time for 6th graders. Given that these variables were measured with only one or two items, these insignificant findings should be interpreted with caution. Until these factors are examined in a more systematic and comprehensive fashion, they should not be dismissed as potential correlates of bullying behavior. The findings from this study have implications for prevention programs. First, messages that teachers and other significant adults in the community convey to students appear to have an important impact on students’ behavior. Adults need to be mindful of the messages they are giving about the use of violence and teasing in their everyday interactions. Second, the significance of two factors—beliefs supportive of violence and confidence in using nonviolent strategies on future bullying behavior—suggests a role for teaching and practicing prosocial skills. However, as teaching social skills alone may not be effective in reducing violence (Dusenbury, Falco, Lake, Brannigan, & Bosworth, 1997; Tolan & Guerra, 1994), research is needed to explore how this instruction can be coupled with other prevention strategies, such as changing the school environment or providing teacher and parent training to reduce bullying behavior. Although this study contributes to a better understanding of potential correlates of bullying, the study design has several limitations that should be considered. First, the data were obtained from self-reports. Second, bullying was measured in terms of behavior during the past 30 days. Thus, the systematic or chronic nature of bullying behaviors was not assessed. Third, the research was conducted with a sample that consisted primarily of White students, with relatively few minorities, and the age range of participants was narrow. In addition, although the study sample matched characteristics of the larger school, the participation rate was less than 50 percent due to the need to secure active parental consent. As a result, it is possible that the students at the greatest risk for bullying may be underrepresented in the sample. These study design problems limit our ability to general-

28

D. L. Espelage et al.

ize the results of this research to a more ethnically diverse population from a different age group and the results may not generalize to more aggressive youth. With these limitations in mind, this study raises several important questions about the stability of bullying behaviors among young adolescents. Clearly, more research needs to be done to better predict bullying during the early middle-school years. The prospective correlates of bullying during the 6th grade identified in this study may prove useful to those who are designing prevention programs.

REFERENCES Begley, S. (1999, May 3). Why the young kill. Newsweek, 133 (18), 32–35. Bentley, K. M., & Li, A. K. F. (1995). Bully and victim problems in elementary schools and students’ beliefs about aggression. Canadian Journal of School Psychology, 11, 153–165. Bosworth, K., Espelage, D., & Simon, T. (1999). Factors associated with bullying behavior among early adolescents. Journal of Early Adolescence, 19, 341–362. Bosworth, K., Espelage, D., DuBay, T., Dahlberg, L., & Daytner, G. (1996). Using multimedia to teach conflict resolution skills to young adolescents. American Journal of Preventive Medicine, 12, 229–255. Bosworth, K., Espelage, D. L., Daytner, G., DuBay, T., & Karageorge, K. (2000). The effectiveness of a multimedia violence prevention program for early adolescents. American Journal of Health Behavior, 24, 268–280. Boulton, M. J., & Smith, P. K. (1994). Bully/victim problems in middle-school children: Stability, self-perceived competence, peer perceptions, and peer acceptance. British Journal of Developmental Psychology, 12, 315–329. Boulton, M. J., & Underwood, K. (1992). Bully/victim problems among middle-school children. British Journal of Educational Psychology, 62, 73–87. Cairns, R. B., Cairns, B. D., Neckerman, H. J., Ferguson, L. L., & Garipy, J. L. (1989). Growth and aggression: 1. Childhood to early adolescence. Developmental Psychology, 25, 320–330. Corsaro, W. A., & Eder, D. (1990). Children’s peer cultures. Annual Review of Sociology, 16, 197–220. Crockett, L. J., Losoff, M., & Petersen, A. C. (1984). Perceptions of the peer group and friendship in early adolescence. Journal of Early Adolescence, 4, 155–181. Dahlberg, L. L., Toal, S. B., & Behrens, C. B. (1998). Measuring violence-related attitudes, beliefs, and behaviors among youths: A compendium of assessment tools. Atlanta, GA: Centers for Disease Control and Prevention, National Center for Injury Prevention and Control. 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. Dornbusch, S. M. (1989). The sociology of adolescence. Annual Review of Sociology, 15, 233–259. Dumas, J. E., Neese, D. E., Prinz, R. J., & Blechman, E. A. (1996). Short-term stability of aggression, peer rejection, and depressive symptoms in middle childhood. Journal of Abnormal Child Psychology, 24, 105–119. Dusenbury, L., Falco, M., Lake, A., Brannigan, R., & Bosworth, K. (1997). Nine critical elements of promising violence prevention programs. Journal of School Health, 67, 409–414. Eder, D. (1985). The cycle of popularity: Interpersonal relations among female adolescents. Sociology of Education, 58, 154–165. Eder, D. (1995). School talk: Gender and adolescent culture. New Brunswick, New Jersey: Rutgers University Press. Espelage, D. L., & Holt, M. (in press). Bullying and victimization during early adolescence: Peer influences and psychosocial correlates. Journal of Emotional Abuse.

Stability of Bullying in Middle School29 Espelage, D. L., Bosworth, K., & Simon, T. (2000). Examining the social environment of middleschool students who bully. Journal of Counseling and Development, 78, 326–333. Farrington, D. P. (1991). Childhood aggression and adult violence: Early precursors and later-life outcomes. In D. J. Pepler & K. H. Rubin (Eds.), The development and treatment of childhood aggression (pp. 5–29). Hillsdale, NJ: Erlbaum. Goodenow, C. (1993). The psychological sense of school membership among adolescents: Scale development and educational correlates. Psychology in the Schools, 30, 79–90. Gorman-Smith, D., &Tolan, P. H. (1998). The role of exposure to community violence and developmental problems among inner-city youth. Development and Psychopathology, 10, 101–116. Gorman-Smith, D., Tolan, P. H., Zelli, A., & Huesmann, L. R. (1996). The relation of family functioning to violence among inner-city minority youths. Journal of Family Psychology, 10, 115–129. Hoover, J. H., & Oliver, R. (1996). The bullying prevention handbook: A guide for principals, teachers, and counselors. Bloomington, IN: National Educational Service. Hoover, J. H., Oliver, R., & Thomson, K. (1993). Perceived victimization by school bullies: New research and future directions. Journal of Humanistic Education and Development, 32, 76–84. Hoover, J. H., Oliver, R., & Hazier, R. J. (1992). Bullying: Perceptions of adolescent victims in the midwestern USA. School Psychology International, 13, 5–16. Huesmann, L. R., Eron, L. D., Lefkowitz, M. M., & Walder, L. O. (1984). Stability of aggression over time and generations. Developmental Psychology, 20, 1120–1234. Klicpera, C, & Klicpera, B. G. (1996) The situation of bullies and victims of aggressive acts in school. Praxis der Kinderpsychologie und Kinderpsychiatrie, 45, 2–9. Limber, S. P., Cummingham, P., Florx, V., Ivey, J., Nation, M., Chai, S., & Melton, G. (1997, June/ July). Bullying among school children: Preliminary findings from a school-based intervention program. Paper presented at the Fifth International Family Violence Research Conference, Durham, NC. 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. Loeber, R., & Stouthamer-Loeber, M. (1998). Development of juvenile aggression and violence: Some common misconceptions and controversies. American Psychologist, 53, 242–259. Moffitt, T. E., Caspi, A., Dickson, N., Silva, P. & Stanton, W. (1996). Childhood-onset versus adolescent-onset antisocial conduct problems in males: Natural history from ages 3 to 18 years. Development and Psychopathology, 8, 399–424. Olweus, D. (1979). Stability of aggressive reaction patterns in males: A review. Psychological Bulletin, 86, 852–875. Olweus, D. (1994). Bullying at school: Long-term outcomes for the victims and an effective schoolbased intervention program. In L. R. Huesmann, Aggressive behavior: Current perspectives (pp. 97–130). New York: Plenum. Patterson, G. R., Reid, J. B., & Dishion, T. J. (1992). A social interactional approach: IV. Antisocial boys. Eugene, OR: Castalia. Pellegrini, A. D., Bartini, M., & Brooks, F. (1999). School bullies, victims, and aggressive victims: Factors relating to group affiliation and victimization in early adolescence. Journal of Educational Psychology, 91, 216–224. Pope, A. W., & Bierman, K. L. (1999). Predicting adolescent peer problems and antisocial activities: The relative roles of aggression and dysregulation. Developmental Psychology, 35, 335–346. Rigby, K., Cox, I., & Black, G. (1997). Cooperativeness and bully/victim problems among Australian schoolchildren. The Journal of Social Psychology, 137, 357–368. Rigby, K., & Slee, P. (1999). Suicidal ideation among adolescent school children, involvement in bully-victim problems, and perceived social support. Suicide and Life-Threatening Behavior, 29, 119–130. 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.

30

D. L. Espelage et al.

Salmivalli, C, Lagerspetz, K., Bjorkqvist, K., Osterman, K., & Kaukiainen, A. (1996). Bullying as a group process: Participant roles and their relations to social status within the group. Aggressive Behavior, 22, 1–15. Sebald, H. (1992). Adolescence. Upper Saddle River. NJ: Prentice Hall, Inc. Slee, P. T. (1995). Peer victimization and its relationship to depression among Australian primary school students. Personality & Individual Differences, 18, 57–62. Slee, P. T., & Rigby, K. (1993). The relationship of Eysenck’s personality factors and self-esteem to bully-victim behaviour in Australia. Personality & Individual Differences,14, 371–373. Thompson, D., & Sharp, S. (1998). The dynamics of victimisation and rejection in school. In D. Thompson & S. Sharp (Eds.), Improving schools: Establishing and integrating whole school behavior policies (pp. 11–25). London: David Fulton. Thornberry, T. P. (1994). Violent families and youth violence. Office of Juvénile Justice and Delinquency Prevention (Fact Sheet No. 21, 1994). Washington, DC: Department of Justice. Tomada, G., & Schneider, B. H. (1997). Relational aggression, gender, and peer acceptance: Invariance across culture, stability over time, and concordance among informants. Developmental Psychology, 33, 601–609. Tolan, P. H., Cromwell, R. E., & Braswell, M. (1986). The application of family therapy to juvenile delinquency: A critical review of the literature. Family Process, 15, 619–649. Tolan, P., & Guerra, N. (1994). What works in reducing adolescent violence: An empirical review of the field. The Center for the Study and Prevention of Violence, Institute for Behavioral Sciences, University of Colorado, Boulder, CO. Webster, D. W, Gainer, P., & Champion, H. R. (1993). Weapon carrying among inner-city junior high school students: Defensive behavior vs. aggressive delinquency. American Journal of Public Health, 83, 1604–1608. Youniss, J., & Smollar, J. (1985). Adolescent relations with mothers, fathers, and friends. Chicago: University of Chicago Press. Zumkley, H. (1994). The stability of aggressive behavior: Ameta-analysis. German Journal of Psychology, 18, 273–281. Acknowledgment. This research was supported by the Centers for Disease Control and Prevention, National Center for Injury Prevention and Control (Grant No. U81/CCU510049-03). Offprints. Requests for offprints should be directed to Dorothy L. Espelage, PhD, Department of Educational Psychology, University of Illinois, Urbana-Champaign, College of Education, 226 Education Building, 1310 South Sixth Street, Champaign, IL 61820-6990.

Violence and Victims, Volume 30, Number 3, 2015

Understanding Ecological Factors Associated With Bullying Across the Elementary to Middle School Transition in the United States Dorothy L. Espelage, PhD University of Illinois at Urbana-Champaign

Jun Sung Hong, PhD Wayne State University, Michigan

Mrinalini A. Rao, PhD Yale University, Connecticut

Robert Thornberg, PhD Linköping University, Sweden This study examines sociodemographic characteristics and social-environmental factors associated with bullying during the elementary to middle school transition from a sample of 5th-grade students (n 5 300) in 3 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

E

arly 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

© 2015 Springer Publishing Company31

32

Espelage et al. 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, Vivolo-Kantor, Hamburger, & Lumpkin, 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 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 sociodemographic characteristics and social-environmental factors (i.e., relationships with parents, peers, and teachers and sense of school belonging) predict changes in bullying during the elementary to middle 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.

Sociodemographics Researchers have examined the influence of sociodemographic variables, particularly biological 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, 2014; 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, whereas boys engage in multiple forms of aggression (Neal, 2007). Despite these findings, several studies have failed to document significant biological sex differences in relational aggression or social forms of bullying (Card, Stucky, Sawalani, & Little, 2008; Crick, Casas, & Mosher, 1997). In addition to biological 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 (Espelage, Basile, & Hamburger, 2012; Low & Espelage, 2012), these studies have yielded small effect sizes. Thus, the research on both biological sex and race/ethnicity differences in reports of bullying perpetration are inconsistent and limited. Thus, we include these variables in our models.

Bullying Across Elementary to Middle School33

Caregivers According to social learning theory and social interaction learning theory in particular, one can assert that maladaptive and aggressive social interactions with peers originate 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 et al., 2014). In contrast, when bullying victims have warm relationships with their families, they have more positive outcomes, both emotionally and behaviorally (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 to 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 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 (C. 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 to 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, Roberts, Zeisel, & Rowley, 2008) because of less structure and supervision in middle schools (Vaillancourt et al., 2010). 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 to middle school transition.

School Belongingness Students’ sense of school belonging declines during the middle school years (Anderman, 2003); consequently, bullying increases from elementary to middle school (Pellegrini, 2002; Pellegrini & Long, 2002). Moreover, as bullying increases, school connectedness tends to

34

Espelage et al.

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 to 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 5 10.83 years, SD 5 0.52). Of these, 158 (53%) were males and 142 (47%) were females. Racial/ethnic composition included 132 (44%) European Americans, 39 (13%) African Americans, 61 (20%) Hispanic/Latino Americans, 53 (18%) Asian Americans, and 15 (5%) indicated Biracial/Multiracial. Students receiving free/reduced lunch ranged from 25% to 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 5 11.32 years, SD 5 0.53). 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 or daughter to participate. Response rates ranged from 95% to 97% among the three schools. Students gave their assent by signing the survey coversheet.

Bullying Across Elementary to Middle School35

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 3 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 45-min 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 T1 and at T2 as the input variables to determine groups of youth who reported no change, increases, or decreases in bullying perpetration across the elementary to middle school transition. Third, when a cluster solution was determined, these clusters were evaluated as a function of biological sex or race/ethnicity. Finally, cluster differences on students self-reports of their family, peer, teacher, and school experiences are presented.

Measures Sociodemographic Scale. Participants completed a sociodemographic questionnaire that included questions about their biological sex, race/ethnicity, and age. Bullying and Victimization Scales. Self-Reported Bullying. We used the 9-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 moderately with the Youth Self-Report (YSR) Aggression Scale (r 5 .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 5 3.50, SD 5 6.50) from their peers than students who did not self-report high levels of bullying perpetration (M 5 0.98; SD 5 1.10; Espelage et al., 2003). This scale is also distinct from pure aggression in factor analyses (Espelage et al., 2014). Acceptable estimates of the scale’s internal consistency were found for the current sample (a 5 .90 for T1, a 5 .85 for T2). Self-Reported Victimization. We used the 4-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 (a 5 .88 for T1, a 5 .87 for T2).

36

Espelage et al.

Self-Perception Scales. The Self-Perception Profile for Children (SPPC; Harter, 1985) provides a self-report self-assessment of 8–13-year-old students. The current investigation only included items corresponding to the Global Self-Worth and Social Acceptance subscales because these domains are shown to be important in the elementary to middle school transition (Robinson, Garber, & Hilsman, 1995). A sample item from Global 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 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-Worth: a 5 .74 for T1, a 5 .75 for T2; Social Acceptance: a 5 .74 for T1, a 5 .75 for T2). Adult Support Scales. We used the People in My Life (PIML) scale (E. Cook, Greenberg, & Kusche, 1995) to assess participants’ relationships with adults. PIML is an adaptation of the Inventory of Parent and Peer Attachment (Armsden & Greenberg, 1987) and was created to assess 10–12-year-old participants’ attachments to their parents and teachers. Relationship With Caregivers. PIML (E. Cook et al., 1995) was used, which included three subscales (Trust, Communication, and Alienation). Trust subscale contains 10 items (e.g., “I trust my parents.”), Communication subscale consists of 5 items (e.g., “My parents listen to what I have to say.”), and Alienation subscale includes 5 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 5 almost never or never true; 4 5 almost always or always true). Acceptable estimates of internal consistency for the Trust (a 5 .84 for T1, a 5 .88 for T2), Communication (a 5 .72 for T1, a 5 .70 for T2), and Alienation (a 5 .65 for T1, a 5 .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 as follows: “My teachers are proud of the things I do” (Affiliation With Teachers subscale, eight items) and “I get upset easily with my teachers” (Dissatisfaction With Teachers subscale, three 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 5 almost never or never true; 4 5 almost always or always true). We found acceptable estimates of internal consistency for the sample for both Affiliation With Teachers subscale (a 5 .90 for T1, a 5 .89 for T2) and Dissatisfaction With Teachers subscale (a 5 .76 for T1, a 5 .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 5 not at all true; 5 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 (a 5 .84 for T1, a 5 .87 for T2).

Bullying Across Elementary to Middle School37

Internalized Symptoms Scales. We used the 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 data of 11–18-year-old samples (Achenbach, 1991). Further, these symptoms are found to be common among students during the elementary to middle school transition (Seidman, Allen, Aber, Mitchell, & Feinmann, 1994). The Withdrawn subscale consists of 5 items (e.g., “I would rather be alone than with others.”), the Somatic Complaints subscale includes 9 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 (a 5 .64 for T1, a 5 .65 for T2), Somatic Complaints (a 5 .70 for T1, a 5 .75 for T2), and Anxious/Depressed (a 5 .86 for T1, a 5 .87 for T2) subscales.

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 several 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 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 To assess different bully perpetration trajectories across the transition, we conducted k-means cluster analysis to create bully-change subtypes. Items from University of Illinois Bullying Scale 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 using both of these methods suggested that a five-cluster solution was appropriate for the data.

38

TABLE 1.  Means, Standard Deviations, and Bivariate Correlations for Variables at Time 1 Variable Social acceptance Global self-worth

1

2

3

4

5

6

7

8

9

10

11

12

13

(.77) .35***

(.74)

Primary caregiver  Trust

.23**

.22**

 Communication

.18**

.18**

 Alienation Teacher affiliation Teacher dissatisfaction School belongingness

(.84) .73***

(.72)

2.08

2.13

2.47***

2.42***

.11

.13

.34***

.33***

2.03

2.06

.43***

.30**

2.17* .31***

2.10 .26***

(.65) 2.18**

(.90)

.22**

2.43***

2.22**

.44***

(.76) 2.20**

(.84)

Depression/ anxiety

2.29***

2.50***

2.19**

2.10

.33***

2.08

.12

2.34***

Withdrawn

2.36***

2.38***

2.24***

2.17**

.31***

2.10

.13*

2.31***

.70***

Somatic complaints

2.12

2.32***

2.07

2.03

.23***

2.08

.19**

2.18**

.61***

.40***

2.22**

2.15*

2.10

.13

2.25***

.13*

2.24***

.31***

.20**

.18**

2.36***

2.25***

2.04

.18**

2.16*

.16*

2.36***

.56***

.44***

.37***

Bullying Victimization

.07 2.32***

(.86) (.64) (.70) (.82) .35***

(.88)

2.89

3.09

3.50

3.04

1.71

3.08

1.69

3.55

1.50

1.59

1.50

1.64

2.07

SD

0.71

0.69

0.49

0.66

0.57

0.77

0.78

0.69

0.37

0.37

0.34

0.66

1.12

Note. Internal consistencies for all scales are reported along the diagonal in parentheses. *p , .05. **p , .01. ***p , .001.

Espelage et al.

M (N 5 237)

TABLE 2.  Means, Standard Deviations, and Bivariate Correlations for Variables at Time 2

Social acceptance Global self-worth

1

2

3

4

5

6

7

8

9

10

11

12

13

(.78) .48***

(.75)

Primary caregiver  Trust

.17**

.44***

 Communication

.15*

.36***

.74***

2.32***

2.47***

2.38***

.37***

.32***

 Alienation

2.12

Teacher affiliation

2.07

.10

Teacher dissatisfaction

2.08

2.08

School belongingness

(.88)

2.21**

(.78)

2.18**

(.70) 2.19** .18**

(.89) 2.48***

.32***

.36***

.40***

.34***

2.32***

Depression/ anxiety

2.32***

2.55***

2.40***

2.29***

.31***

2.15*

.08

2.40***

Withdrawn

2.36***

2.41***

2.29***

2.28***

.26***

2.09

.08

2.40***

.74***

Somatic complaints

2.20**

2.29***

2.17**

2.10

.16**

2.07

.02

2.20**

.62***

.46*** (.75)

2.09

2.27***

2.10

.31***

2.38***

.32***

2.43***

.46***

.10

.10

2.33***

2.22**

2.17**

.13*

2.02

.10

2.31***

.23**

.37***

.36***

Bullying Victimization

.10 2.29***

.49***

(.79) 2.29**

(.87) (.87) (.64)

(.86) .24***

(.87)

M (N 5 237)

3.05

3.15

3.48

3.10

1.76

2.98

1.72

3.60

1.68

1.52

1.38

1.64

1.79

SD

0.65

0.65

0.55

0.90

0.61

0.81

0.84

0.71

0.35

0.36

0.34

0.82

0.97

Note. Internal consistencies for all scales are reported along the diagonal in parentheses. *p , .05. **p , .01. ***p , .001.

Bullying Across Elementary to Middle School39

Variable

40

Espelage et al.

Following calculation of the cluster solution via these hierarchical methods, the data were then reanalyzed 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.

Initial Cluster Descriptions Cluster 1 (C1), defined as the Stable Low cluster, included most participants (N 5 146, 61.6%; see Figure 1). These students had both the lowest mean bullying scale score at T1 (M 5 1.27, SD 5 0.23) and T2 (M 5 1.24, SD 5 0.28). Cluster 2 (C2), defined as Decreasing, included participants (N 5 44, 18.6%) whose bullying decreased over the transition, with moderate levels of Bullying (T1; M 5 2.32, SD 5 0.53) and decreased levels of Bullying (T2; M 5 1.61, SD 5 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 5 2.35, SD 5 0.67) and a higher score at T2 (M 5 2.65, SD 5 0.26). A paired-sample t test indicated that this increase was significant (p , .05). The fourth cluster (C4), High Increasing, included participants (N 5 11, 4.6%) who had low bullying scores at T1 (M 5 1.79, SD 5 0.57) and high scores at T2 (M 5 3.85, SD 5 0.60). Finally, Cluster 5 (C5) was designated the Stable High group (N 5 2, 0.8%) and included participants who had the highest bullying scores at T1 (M 5 4.00, SD 5 0.63) and T2 (M 5 4.78, SD 5 0.16). Differences among the clusters on Bullying at T1, F(4, 223) 5 104.04, p , .001, and T2, F(4, 223) 5 324.56, p , .001, scores were significant. Further, post hoc tests revealed significant differences between all cluster pairings.

6 Stable low 5 Decreasing

4 3

Moderately increasing

2

High increasing

1 Stable high 0 Mean bully scale (Time 1)

Mean bully scale (Time 2)

Figure 1.  Means of Bully Perpetration Scale (Time 1 and Time 2) for Cluster Membership.

Bullying Across Elementary to Middle School41

Final Cluster Descriptions Participants in C5 were outliers and were removed from subsequent analyses. Further, because our goal was to explore increases or decreases in bullying over the transition, C3 and C4 were combined as both groups. We found that bullying increased significantly from T1 to T2. Thus, in the subsequent analyses, three clusters were used. C1, labeled as Stable Low (N 5 146, 61.6%), included participants who reported the lowest levels of bullying at both T1 and 2. C2, labeled as Decreasing (N 5 44, 18.6%), included participants whose bullying decreased over the transition. Finally, C3, Increasing, (N 5 36, 15.2%) included participants with moderate Bullying at T1 (M 5 2.18, SD 5 0.68) and higher scores at T2 (M 5 3.02, SD 5 0.68).

Cluster Sociodemographic Characteristics Results revealed significant differences for biological sex, x2(2, N 5 226) 5 8.63, p , .05, and for race/ethnicity, x2(2, N 5 226) 5 56.41, p , .001. With respect to biological sex, a greater percentage of females (73.5%) than males (55.0%) were in the Stable Low cluster, whereas a greater percentage of males were in the two remaining groups (25.7% male vs. 13.7% female: Decreasing cluster; 19.3% male vs. 12.8% female: Increasing cluster). Seventy percent of European Americans, 74% of Hispanic/Latino Americans, and 66% of Asian Americans were in the Stable Low group compared to 26% of African Americans. In addition, 49% of African Americans were in the Increasing group compared to 11% of European Americans, 8% of Hispanic/Latino Americans, and 7% of Asian Americans. Finally, 19% of European Americans, 11% of African Americans, and 18% of Hispanic/ Latino Americans and Asian Americans were classified in the Decreasing group. Given that clusters differed by sociodemographic characteristics, in subsequent analyses, biological sex and race/ethnicity were included in interaction terms. Cluster Group Membership and Changes in Adult Support, Social Acceptance, Psychosocial Functioning, and School Belonging. One overall repeated measures MANOVA was conducted to examine differences among the three clusters on the variables of interest: three caregiver scales, two teacher scales, school sense of belonging, global self-worth, social acceptance, peer victimization, depression/anxiety, withdrawn behaviors, and somatization. The repeated measure was Time with two levels (T1 and T2). Cluster Membership was treated as an independent variable. Thus, the critical interaction between Time and Cluster Membership was of interest for this study. We found an overall MANOVA effect for Cluster Membership by Time on the scales (F 5 3.31, p , .001, h2 5 .17; see Table 3). The overall MANOVA effect for the three-way interaction among Cluster Membership, Time, and Biological Sex was not significant. Significant multivariate effects were followed with ANOVAs, and significant ANOVAs were followed by Tukey post hoc comparisons. A final MANOVA was calculated to include race/ethnicity in addition to Cluster Membership, Time, and Biological Sex. The results indicated that the three-way interaction among Time, Cluster Membership, and Race/Ethnicity was not significant, indicating that the change in bullying over the transition did not vary as a factor of race/ethnicity. In addition, the fourway interaction among Time, Cluster Membership, Biological Sex, and Race/Ethnicity was not significant. Cluster by Time Influences on Primary Caregiver Relationships. A significant ANOVA effect was found for the critical interaction of Cluster Membership 3 Time on Primary Caregiver Trust, F(1, 236) 5 5.46, p , .01, h2 5 .05. Tukey post hoc comparisons

42

TABLE 3.  Means and Standard Deviations on Variables for Cluster by Time Cluster 1: No Change (N 5 146) T1

Cluster 2: Decreasing (N 5 44)

Cluster 3: Increasing (N 5 36)

T1

T1

T2

T2

T2

Cluster 3 Time

M

SD

M

SD

M

SD

M

SD

M

SD

M

SD

F

h2

 Trust

3.56

0.47

3.55

0.50

3.46

0.49

3.53

0.42

3.48

0.43

3.21

0.72

5.46***

.05

 Communication

3.09

0.68

3.15

0.68

3.02

0.53

3.19

0.55

3.09

0.65

2.93

0.85

2.32

.02

 Alienation

1.69

0.58

1.70

0.56

1.76

0.51

1.71

0.53

1.69

0.57

2.04

0.64

5.86**

.02

Teacher affiliation

3.13

0.76

3.12

0.74

2.97

0.68

2.90

0.69

2.88

0.81

2.43

0.91

3.49*

.03

Teacher dissatisfaction

1.62

0.74

1.58

0.73

1.85

0.80

1.63

0.62

1.91

0.96

2.26

1.14

2.96*

.03

School belongingness

3.62

0.68

3.77

0.68

3.46

0.50

3.56

0.57

3.57

0.73

3.04

0.58

12.31***

.11

Depression/ anxiety

1.45

0.33

1.33

0.32

1.61

0.35

1.40

0.26

1.58

0.41

1.58

0.46

4.79**

.05

Withdrawn

1.53

0.33

1.47

0.37

1.62

0.36

1.56

0.35

1.67

0.48

1.59

0.37

0.05

.00

Somatic complaints

1.46

0.33

1.36

0.33

1.56

0.31

1.40

0.31

1.63

0.37

1.53

0.34

0.71

.01

Victimization

1.83

0.94

1.62

0.81

2.49

1.04

1.72

0.72

2.41

1.44

2.34

1.29

6.84**

.06

Global self-worth

3.18

0.63

3.22

0.62

2.97

0.75

3.03

0.48

2.80

0.81

3.06

0.82

1.16

.01

Social acceptance

2.83

0.70

3.06

0.60

3.01

0.67

3.09

0.54

2.96

0.74

3.14

0.68

0.74

.01

Variable Primary caregiver

Espelage et al.

*p , .05. **p , .01. ***p , .001.

Bullying Across Elementary to Middle School43

indicated a significant difference between C1 (Stable Low) and C3 (Increasing), which suggests that participants with increased bullying over the transition had less trust in their caregivers at T2 compared to those whose bullying level was low over the transition. Cluster Membership 3 Time interaction was not significant for Primary Caregiver Communication or Alienation. Cluster by Time Influences on Teacher Relationships. Significant ANOVA effects were found for Cluster Membership 3 Time interaction on both Teacher Affiliation, F(1, 236) 5 3.49, p , .05, h2 5 .03, and Teacher Dissatisfaction, F(1, 236) 5 2.96, p , .01, h2 5 .03. Tukey post hoc comparisons indicated a significant difference between C1 (Stable Low) and C3 (Increasing) for Teacher Affiliation, which suggests that participants with increases in Bullying (T2) reported a decrease in teacher affiliation over the transition. For Teacher Dissatisfaction, Tukey post hoc comparisons indicated significant differences between the Increasing group and the other two clusters. Participants in C1 (Stable Low) and C2 (Decreasing) both experienced decreases in Teacher Dissatisfaction, whereas those in C2 (Increasing) reported higher levels of Teacher Dissatisfaction across the transition. Cluster by Time Influences on School Sense of Belonging. A significant ANOVA effect was found for the interaction of Cluster Membership 3 Time on Psychological Sense of School Membership, F(1, 236) 5 12.31, p , .001, h2 5 .11. Tukey post hoc comparisons indicated there a significant difference between C1 (Stable Low) and C3 (Increasing), suggesting that participants with increases in levels of bullying reported a decrease in School Sense of Belonging. Cluster by Time Influences on Internalizing Symptoms. A significant ANOVA effect was found for the interaction of Cluster Membership 3 Time on Depression/ Anxiety, F(1, 236) 5 2.96, p , .01, h2 5 .05. Tukey post hoc comparisons indicated a significant difference between C1 (Stable Low) and C3 (Increasers), which suggests that participants whose level of bullying increased reported increases in Depression/Anxiety from T1 to T2, compared to students whose bullying level was low at both time points. Cluster Membership 3 Time interaction was not significant for Withdrawn Behaviors or Somatic Complaints. Cluster by Time Influences on Peer Victimization. A significant ANOVA effect was found for the critical interaction of Cluster Membership 3 Time on Victimization, F(1, 236) 5 3.49, p , .05, h2 5 .03. Tukey post hoc comparisons indicated that participants in C1 differed significantly from those in the other two clusters. Participants with decreasing levels of bullying reported less victimization in middle school than in elementary school. Although the mean scores on victimization decreased for all students over the transition, participants with an increasing level of bullying reported the highest levels of Victimization. Cluster by Time Influences on Global Self-Worth and on Social Acceptance. Cluster 3 Time was not significant for Global Self-Worth or Social Acceptance, which suggests that Cluster Membership did not influence changes in participants’ reported levels of SelfWorth or Social Acceptance during the transition.

DISCUSSION Although bullying increases in middle school (Pellegrini, 2002; Pellegrini & Long, 2002), relatively little is known about changes across the elementary to middle school transition. Results here indicate that bullying during this transition appear to be heterogeneous.

44

Espelage et al.

Although several participants reported an increase, most reported stable low levels of bullying in both contexts. However, one cluster of participants showed a major decrease in their self-reported bullying. Nevertheless, our findings highlight the importance of an ecological approach in identifying individual and social-environmental factors during the transitional period. We also found that over the transition, increases in self-reported bullying were associated with several social-environmental variables. As the students transitioned to middle school, those who reported increases in bullying also reported decreases in trust in their caregivers and increases in negative relationships with their teachers. It is more challenging to develop supportive teacher–student relationships in middle school than in elementary school because students change classrooms and spend less time with a single teacher. However, building positive, supportive, and trusting relationships between teachers and students appears to be crucial in middle schools, as evidenced by research findings (e.g., Barboza et al., 2009). Moreover, increases in bullying were also associated with higher rates of depression/ anxiety in the present findings. Hence, efforts in promoting students’ psychological health and well-being should be included in bullying prevention and intervention, not only to support the victims (Ttofi, Farrington, Lösel, & Loeber, 2011) but also to change the behaviors of bullies. The largest effect was found for school belongingness; that is, increases in bullying were associated with decreases in school belongingness. Middle school transition can be stressful, putting many students at risk for negative outcomes (Prinstein & Aikins, 2004). Thus, adults need to focus on promoting a positive school climate (Bandyopadhyay, Cornell, & Konold, 2009; Barboza et al., 2009; C. Cook et al., 2010; Gendron, Williams, & Guerra, 2011), which is characterized as having positive teacher–student relationships (Barboza et al., 2009; Bergin & Bergin, 2009; Jennings & Greenberg, 2009; Meehan et al., 2003) and a strong sense of school belonging among the students (Bergin & Bergin, 2009; Tillery, Varjas, Roach, Kuperminc, & Meyers, 2013). We should also note that the relationship between school climate and bullying is bidirectional; that is, a negative school climate can reinforce bullying behavior among students, but the presence of bullying can also reinforce negative perceptions of the school climate. Nevertheless, Anderman’s (2003) longitudinal study found that although middle school students on average perceived a declining sense of school belonging over time, students who reported that their teachers promoted mutual respect in classes also reported less negative change in school belongingness over time. Our findings also suggest that students who reported an increasing level of bullying also reported the highest levels of victimization. Researchers in recent years have focused on bully–victim subgroups (i.e., students involved in bullying and victimization) and found that bully victims are at risk for negative outcomes (C. Cook et al., 2010). These students typically engage in bullying as a response to victimization. Another interesting finding was that perceived social acceptance did not change over time as a result of cluster membership. That is, all three clusters reported an increase in social acceptance across the transition to middle school. The finding that students who reported an increase in bullying perpetration, but no reduction in social acceptance, is consistent with a meta-analytic finding that bullying is associated with greater social status during early adolescence (C. Cook et al., 2010). Among the strengths of this study is the short-term longitudinal design. Assessing bullying, adult attachment, school variables, and psychological functioning at the end of fifth grade and again a few months into sixth grade allowed for the isolation of transitional

Bullying Across Elementary to Middle School45

outcomes. This allows for the interpretation of the results that negative experiences in transition are associated with increases in bullying. However, this study also has limitations. These data were all self-reported and did not include data from teachers or caregivers. Thus, some associations that we found might be stronger as an effect of shared method variance. Also, two subscales have low alphas (e.g., Internalizing: Withdrawn; Relationship With Parents: Alienation). Furthermore, a relatively small sample size did not allow for a more fine-grained analysis of racial/ethnic differences across the transition, which is critical to evaluate Garcia Coll et al.’s (1996) integrative model of developmental competencies in minority children. Finally, this study used cluster analysis rather than latent profile analysis because only one continuous measure was used to create the cluster groups. Despite these limitations, our study provides interesting and unique insights into social-environmental factors associated with bullying across the elementary to middle school transition. It appears that teacher support and sense of belonging in school might play a critical role during this period.

REFERENCES Achenbach, T. M. (1991). Manual for the Youth Self-Report and 1991 profile. Burlington, VT: University of Vermont, Department of Psychiatry. Aldenderfer, M. S., & Blashfield, R. K. (1984). Cluster analysis. Newbury Park, CA: Sage. Anderman, L. H. (2003). Academic and social perceptions as predictors of change in middle school students’ sense of school belonging. Journal of Experimental Education, 72, 5–22. http:// dx.doi.org/10.1080/00220970309600877 Armsden, G. C., & Greenberg, M. T. (1987). The Inventory of Parent and Peer Attachment: Individual differences and their relationship to psychological well-being in adolescence. Journal of Youth and Adolescence, 16, 427–454. http://dx.doi.org/10.1007/BF02202939 Bandyopadhyay, S., Cornell, D. G., & Konold, T. R. (2009). Validity of three school climate scales to assess bullying, aggressive attitudes, and help seeking. School Psychology Review, 38, 338–355. Barboza, G. E., Schiamberg, L. B., Oehmke, J., Korzeniewski, S. J., Post, L. A., & Heraux, C. G. (2009). Individual characteristics and the multiple contexts of adolescent bullying: An ecological perspective. Journal of Youth and Adolescence, 38, 101–121. http://dx.doi.org/10.1007/ s10964-008-9271-1 Becker, B. E., & Luthar, S. S. (2002). Social-emotional factors affecting achievement outcomes among disadvantaged students: Closing the achievement gap. Educational Psychologist, 37, 197–214. http://dx.doi.org/10.1207/S15326985EP3704_1 Belgrave, F. Z. (2009). African American girls: Reframing perceptions and changing experiences. New York, NY: Springer Publishing. Bergin, C., & Bergin, D. (2009). Attachment in the classroom. Educational Psychology Review, 21, 141–170. http://dx.doi.org/10.1007/s10648-009-9104-0 Borgen, F. H., & Barnett, D. C. (1987). Applying cluster analysis in counseling psychology research. Journal of Counseling Psychology, 34, 456–468. http://dx.doi.org/10.1037/0022-0167.34.4.456 Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and design. Cambridge, MA: Harvard University Press. Bukowski, W. M., Sippola, L. K., & Newcomb, A. F. (2000). Variations in patterns of attraction of same- and other-sex peers during early adolescence. Developmental Psychology, 36, 147–154. http://dx.doi.org/10.1037/0012-1649.36.2.147 Burchinal, M. R., Roberts, J. E., Zeisel, S. A., & Rowley, S. J. (2008). Social risk and protective factors for African American children’s academic achievement and adjustment during

46

Espelage et al.

the transition to middle school. Developmental Psychology, 44, 286–292. http://dx.doi .org/10.1037/0012-1649.44.1.286 Card, N., Stucky, B., Sawalani, G., & Little, T. (2008). Direct and indirect aggression during childhood and adolescence: A meta-analytic review of gender differences, intercorrelations, and relations to maladjustment. Child Development, 79, 1185–1229. http://dx.doi.org/10.1111/ j.1467-8624.2008.01184.x Cook, C. R., Williams, K. R., Guerra, N. G., Kim, T. E., & Sadek, S. (2010). Predictors of bullying and victimization in childhood and adolescence: A meta-analytic investigation. School Psychology Quarterly, 25, 65–83. http://dx.doi.org/10.1037/a0020149 Cook, E. T., Greenberg, M. T., & Kusche, C. A. (1995, March). People in my life: Attachment relationships in middle childhood. Paper presented at the Society for Research in Child Development, Indianapolis, IN. Crick, N. R., Casas, J. F., & Mosher, M. (1997). Relational and overt aggression in preschool. Developmental Psychology, 33, 579–588. http://dx.doi.org/10.1037/0012-1649.33.4.579 Espelage, D. L. (2012). Bullying prevention: A research dialogue with Dorothy Espelage. Prevention Researcher, 19, 17–19. Espelage, D. L., Basile, K. C., & Hamburger, M. E. (2012). Bullying experiences and co-occurring sexual violence perpetration among middle school students: Shared and unique risk factors. Journal of Adolescent Health, 50, 60–65. http://dx.doi.org/10.1016/j.jadohealth.2011.07.015 Espelage, D. L., & Holt, M. K. (2001). Bullying and victimization during early adolescence: Peer influences and psychosocial correlates. In R. A. Geffner, M. Loring, & C. Young (Eds.), Bullying behavior: Current issues, research, and interventions (pp. 123–142). New York, NY: Haworth Press. 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. http://dx.doi .org/10.1111/1467-8624.00531 Espelage, D. L., Low, S., Rao, M. A., Hong, J. S., & Little, T. (2014). Family violence, bullying, fighting, and substance use among adolescents: A longitudinal transactional model. Journal of Research on Adolescence, 24, 337–349. http://dx.doi.org/10.1111/jora.12060 Espelage, D. L., & Swearer, S. M. (Eds.). (2011). Bullying in North American schools (2nd ed.). New York, NY: Routledge. Farmer, T. W., Petrin, R. A., Robertson, D. L., Fraser, M. W., Hall, C. M., Day. S. H., & Dadisman, K. (2010). Peer relations of bullies, bully-victims, and victims: The two social worlds of bullying in second-grade classrooms. Elementary School Journal, 110, 364–392. http://dx.doi .org/10.1086/648983 Garcia Coll, C., Lamberty, G., Jenkins, R., McAdoo, H. P., Crnic, K., Wasik, B. H., & Garcia, H. V. (1996). An integrative model for the study of developmental competencies in minority children. Child Development, 67, 1891–1914. http://dx.doi.org/10.1111/j.1467-8624.1996.tb01834.x Gendron, B. P., Williams, K. R., & Guerra, N. G. (2011). An analysis of bullying among students within schools: Estimating the effects of individual normative beliefs, self-esteem, and school climate. Journal of School Violence, 10, 150–164. http://dx.doi.org/10.1080/ 15388220.2010.539166 Gladden, R. M., Vivolo-Kantor, A. M., Hamburger, M. E., & Lumpkin, C. D. (2014). Bullying surveillance among youths: Uniform definitions for public health and recommended data elements, version 1.0. Atlanta, GA: National Center for Injury Prevention and Control, Centers for Disease Control and Prevention, and U.S. Department of Education. Goodenow, C. (1993). The Psychological Sense of School Membership among adolescents: Scale development and educational correlates. Psychology in the Schools, 30, 79–90. http://dx.doi .org/10.1002/1520-6807(199301)30:13.0.CO;2-X Hair, J. F. J., & Black, W. C. (2000). Cluster analysis. In L. G. Grimm & P. R. Yarnold (Eds.), Reading and understanding More multivariate statistics (pp. 147–205). Washington, DC: American Psychological Association.

Bullying Across Elementary to Middle School47 Harter, S. (1985). Manual for the self-perception profile for children. Denver, CO: University of Denver. Holt, M. K., & Espelage, D. L. (2007). Perceived social support among bullies, victims, and bully-victims. Journal of Youth and Adolescence, 36, 984–994. http://dx.doi.org/10.1007/s10964-006-9153-3 Hong, J. S., & Espelage, D. L. (2012). A review of research on bullying and peer victimization in school: An ecological systems analysis. Aggression and Violent Behavior, 17, 311–322. http:// dx.doi.org/10.1016/j.avb.2012.03.003 Jennings, P. A., & Greenberg, M. T. (2009). The prosocial classroom: Teacher social and emotional competence in relation to student and classroom outcomes. Review of Educational Research, 79, 491–525. http://dx.doi.org/10.3102/0034654308325693 Kinder, B. N., Curtiss, G., & Kalichman, S. (1991). Cluster analyses of headache-patient MMPI scores: A cross-validation. Psychological Assessment, 3, 226–231. http://dx.doi .org/10.1037/1040-3590.3.2.226 Koo, D. J., Peguero, A. A., & Shekarkhar, Z. (2012). Gender, immigration, and school victimization. Victims & Offenders, 7, 77–96. http://dx.doi.org/10.1080/15564886.2011.629773 Low, S., & Espelage, D. L. (2012). Differentiating cyber bullying perpetration from other forms of peer aggression: Commonalities across race, individual, and family predictors. Psychology of Violence, 3, 39–52. http://dx.doi.org/10.1037/a0030308 McDougall, P., & Hymel, S. (1998). Moving into middle school: Individual differences in the transition experience. Canadian Journal of Behavioural Science, 30, 108–120. http://dx.doi .org/10.1037/h0085811 Meehan, B. T., Hughes, J. N., & Cavell, T. A. (2003). Teacher–student relationships as compensatory resources for aggressive children. Child Development, 74, 1145–1157. http://dx.doi .org/10.1111/1467-8624.00598 Milligan, G. W. (1980). An examination of the effect of six types of error perturbation on fifteen clustering algorithms. Psychometrika, 45, 325–342. http://dx.doi.org/10.1007/BF02293907 Nansel, T. R., Haynie, D. L., & Simons-Morton, B. G. (2003). The association of bullying and victimization with middle school adjustment. Journal of Applied School Psychology, 19, 45–61. http://dx.doi.org/10.1300/J008v19n02_04 Nansel, T. R., Overpeck, M., Pilla, R. S., Ruan, W. J., 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. http://dx.doi.org/10.1001/jama.285.16.2094 Neal, J. W. (2007). Why social networks matter: A structural approach to the study of relational forms of aggression in middle childhood and adolescence. Child and Youth Care Forum, 36, 195–211. http://dx.doi.org/10.1007/s10566-007-9042-2 Patterson, G. R. (1982). Coercive family process. Eugene, OR: Castalia. Patterson, G. R., Dishion, T. J., & Bank, L. (1984). Family interaction: A process model of deviancy training. Aggressive Behavior, 10, 253–267. http://dx.doi.org/10.1002/1098-2337(1984) 10:33.0.CO;2-2 Pellegrini, A. (2002). Bullying, victimization, and sexual harassment during the transition to middle school. Educational Psychologist, 37, 151–163. http://dx.doi.org/10.1207/S15326985EP3703_2 Pellegrini, A. D., & Bartini, M. (2000). A longitudinal study of bullying, victimization, and peer affiliation during the transition from primary school to middle school. American Educational Research Journal, 37, 699–725. http://dx.doi.org/10.3102/00028312037003699 Pellegrini, A., & Long, J. D. (2002). A longitudinal study of bullying, dominance, and victimization during the transition from primary school through secondary school. British Journal of Developmental Psychology, 20, 259–280. http://dx.doi.org/10.1348/026151002166442 Prinstein, M. J., & Aikins, J. W. (2004). Cognitive moderators of the longitudinal association between peer rejection and adolescent depressive symptoms. Journal of Abnormal Child Psychology, 32, 147–158. http://dx.doi.org/10.1023/B:JACP.0000019767.55592.63 Robinson, N. S., Garber, J., & Hilsman, R. (1995). Cognitions and stress: Direct and moderating effects on depressive versus externalizing symptoms during the junior high school transition.

48

Espelage et al.

Journal of Abnormal Psychology, 104, 453–463. http://dx.doi.org/10.1037/0021-843X .104.3.45 Rodkin, P. C., & Hodges, E. V. E. (2003). Bullies and victims in the peer ecology: Four questions for school service providers and social development research. School Psychology Review, 32, 384–400. Seidman, E., Allen, L., Aber, J., Mitchell, C., & Feinmann, J. (1994). The impact of school transitions in early adolescence on the self-system and perceived social context of poor urban youth. Child Development, 65, 507–522. http://dx.doi.org/10.2307/1131399 Tillery, A. D., Varjas, K., Roach, A. T., Kuperminc, G. P., & Meyers, J. (2013). The importance of adult connections in adolescents’ sense of school belonging: Implications for schools and practitioners. Journal of School Violence, 12, 134–155. http://dx.doi.org/10.1080/15388220.2012.762518 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. http://dx.doi .org/10.1108/17596591111132873 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 unsafe and high bullying areas at school. Canadian Journal of School Psychology, 25, 40–54. Varjas, K., Henrich, C. C., & Meyers, J. (2009). Urban middle school students’ perceptions of bullying, cyberbullying, and school safety. Journal of School Violence, 8, 159–176. http://dx.doi .org/10.1080/15388220802074165 Wang, J., Iannotti, R. J., & Nansel, T. R. (2009). School bullying among adolescents in the United States: Physical, verbal, relational, and cyber. Journal of Adolescent Health, 45, 368–375. http://dx.doi.org/10.1016/j.jadohealth.2009.03.021 Ward, J. H. (1963). Hierarchical grouping to optimize an objective function. Journal of the American Statistical Association, 58, 236–244. http://dx.doi.org/10.1080/01621459.1963.10500845 Young, D. H. (2004). Does school connectedness predict bullying? An analysis of perceptions among public middle school students. Dissertation Abstracts International Selection A: Humanities and Social Sciences, 64, 3959. Author’s Note. Both Jun Sung Hong and Mrinalini A. Rao contributed equally to the article. Correspondence regarding this article should be directed to Dorothy L. Espelage, PhD, Dept. of Educational Psychology, Child Development Division, University of Illinois, 220A Education, 1310 S. Sixth Street, Champaign, IL 61820-6925. E-mail: [email protected]

Violence and Victims, Volume 29, Number 4, 2014

Individual and Social Network Predictors of Physical Bullying: A Longitudinal Study of Taiwanese Early Adolescents Hsi-Sheng Wei, PhD National Taipei University, Taiwan

Wonjae Lee, PhD Graduate School of Culture Technology, Korea Advanced Institute of Science and Technology, South Korea This study followed 125 7th-grade students in Taiwan for the entire school year and analyzed the individual and social network factors predicting their involvement in physical bullying over 5 waves of data. Using self-reports of bullying experiences, 20 classroomlevel networks of bullying and friendship were constructed for 4 classrooms and 5 temporal points, from which 4 individual-level network measures were calculated. They included bully and victim centrality, popularity, and embeddedness in friendship networks. A series of mixed models for repeated measures were constructed to predict students’ bully and victim centrality in bullying network at time t 1 1. Compared to girls, boys were more likely to be both the bullies and victims. Lower self-esteem and higher family economic status contributed to victim centrality. Having parents married and living together predicted lower bully centrality. Higher educational level of parents predicted lower victim and bully centrality. Regarding the social network factors, students’ bully centrality at t positively predicted their bully centrality at t 1 1, whereas victim centrality predicted their subsequent victim centrality. Interaction effects between friendship network and bullying network were observed. Embeddedness in friendship network reduced victim centrality at t 1 1 except for those students with low victim centrality at t. For those with high victim centrality at t, popularity increased their risk of physical victimization over time. Implications for research and practice are discussed.

Keywords: physical bullying; social network; longitudinal; adolescents; Taiwan

P

hysical bullying is a common form of aggression that poses a serious threat to the well-being of students. Many children and adolescents have experienced being hit, kicked, shoved, or other forms of physical victimization in school (e.g., KhouryKassabri, Benbenishty, & Astor, 2005). A large body of research has revealed its prevalence around the world (Craig et al., 2009). A nationally representative survey of 7,182 American students in Grades 6–10 found that 20.8% of the students have physically bullied others or have been bullied at school at least once in the last 2 months (Wang, © 2014 Springer Publishing Company49

50

Wei and Lee

Iannotti, & Nansel, 2009). Another study with a representative population sample of 2,464 Norwegian adolescents aged 12–15 years old showed that 22.5% of the boys and 9.4% of the girls reported physically attacking others during the past 6 months, whereas 14.0% and 6.6% reported being physically assaulted (Undheim & Sund, 2010). Similarly, a recent large-scale, nationally representative survey of Taiwanese adolescents found that 22% of the students aged 12–15 years old had physically bullied others during the past year, whereas 15% reported being physically victimized (Wei, Sun, Liu, & Chen, 2011). These numbers not only highlight the universality of the problem but also call for more preventive efforts by school personnel and educating professionals on how to address this substantial student safety issue (e.g., Jenson, Dieterich, Brisson, Bender, & Powell, 2010). Being bullied has been shown to be associated with various mental health problems, including depression, anxiety, and somatic symptoms (Farrington, Loeber, Stallings, & Ttofi, 2011; Fekkes, Pijpers, Fredriks, Vogels, & Verloove-Vanhorick, 2006; Gladstone, Parker, & Malhi, 2006). It also leads to substance abuse, self-harm behavior, and suicide attempts (Fisher et al., 2012; Klomek, Marrocco, Kleinman, Schonfeld, & Gould, 2007; Radliff, Wheaton, Robinson, & Morris, 2012). Moreover, victims of school bullying often show reduced interest in school and poor academic performance (Boulton, Chau, Whitehand, Amataya, & Murray, 2009; Juvonen, Wang, & Espinoza, 2011; Wei & Williams, 2004). At the same time, school bullying, especially physical bullying (Bender & Lösel, 2011), has been found to be a strong predictor of delinquency and adulthood criminal offending (Farrington et al., 2011; also see Ttofi, Farrington, Lösel, & Loeber, 2011 for a review). It is therefore evident that involvement in bullying either as a victim or a bully has a negative effect on students’ psychosocial well-being and developmental trajectory, which constitutes a significant school health concern for adolescents.

INDIVIDUAL-LEVEL PREDICTORS OF BULLYING To implement effective prevention and intervention measures, it is important for school professionals to identify potential bullies and victims. Previous studies have suggested several risk factors of bullying and victimization. For example, male students are consistently found to be both the perpetrators and targets of physical bullying (e.g., Scheithauer, Hayer, Petermann, & Jugert, 2006). In fact, compared to their female counterparts, males are also generally more involved in other aggressive acts, such as violent crimes (Lauritsen & Heimer, 2008; Lauritsen, Heimer, & Lynch, 2009). Individual characteristics, including low self-esteem and social withdrawal, have been found to be associated with peer victimization (Andreou, 2000; Wei & Chen, 2009). One nationwide study of bullying conducted in Ireland with 8,249 school children aged 8–18 years showed that bullies, victims, or bully/victims all had significantly lower global self-esteem than did students who had neither bullied others nor been bullied (O’Moore & Kirkham, 2001). At the same time, poor academic performance in school was associated with both perpetration and victimization (Hemphill et al., 2012; Tom, Schwartz, Chang, Farver, & Xu, 2010). A reciprocal effect is likely to exist between school violence and school performance because evidence also suggests that children who are bullied in school tend to obtain low levels of academic achievement, particularly when their liking of school is impacted by the victimization experience (Beran, Hughes, & Lupart, 2008). In addition to the previously mentioned individual characteristics, demographic ­variables, such as parent’s education, socioeconomic status (SES), and parental m ­ arriage

Individual and Social Network Predictors of Physical Bullying 51

status, have also been shown to be related to student’s bullying involvement. For example, a recent analysis of 162,305 students from nationally representative samples of 5,998 schools in 35 countries in Europe and North America found that adolescents from less affluent families reported a higher prevalence of being victims of bullying (Due et al., 2009). A longitudinal study of 1,666 seventh- and eighth-grade students from two Korean middle schools found that boys from lower SES families and girls from non-intact families were at an increased risk for bullying (Kim, Boyce, Koh, & Leventhal, 2009). One study of German primary school students showed that low parental educational levels significantly predicted their bullying status as a bully, victim, and bully/victim (Von Marées & Petermann, 2010). Another survey on a representative population sample of 2,464 Norwegian adolescents found that students who bullied or were bullied reported more parental divorce compared to their noninvolved peers (Undheim & Sund, 2010). However, the effects of these demographic variables are less than conclusive, and none of them are based on Taiwanese samples. Empirical investigation is needed to determine the predictive power of these factors on the bullying and victimization among Taiwanese adolescents.

THE SIGNIFICANCE OF SOCIAL NETWORK FACTORS Bullying happens in social contexts and relationships, particularly in the peer context. The bullies and victims often have distinctive peer status and social network positions. Low popularity, peer rejection, and lack of friends are frequently suggested as strong risk factors for victimization (de Bruyn, Cillessen, & Wissink, 2010; Mouttapa, Valente, Gallaher, Rohrbach, & Unger, 2004). Similarly, having more friends and being liked by peers were associated with less victimization (Pellegrini, Bartini, & Brooks, 1999; Wang et al., 2009). Qualitative research has shown that the victims are likely to be marginalized and stigmatized in the peer group (Thornberg, 2011), which makes them increasingly isolated and unable to escape from their negative victim image. From a sociometric perspective, being victimized may contribute to one’s sequent victim status. On the other hand, the bullies are not necessarily rejected by peers (Sentse, Scholte, Salmivalli, & Voeten, 2007). Some longitudinal studies even found that high bullying led to the attainment of high social dominance (e.g., Reijntjes et al., 2013). It is suggested that bullying is one strategy through which the adolescents acquire social status and establish a power hierarchy in the peer group, and this aggressive behavior is expected to decrease as the individual develops satisfactory friendship and network position among the peers (Pellegrini, 2002). In summary, the bullies and victims differ significantly in their peer relation and sociometric status, and these social network factors may predict their ­bullying ­involvement over time. Although recent findings suggest the relevance of peer contexts to school bullying (e.g., Huitsing, Veenstra, Sainio, & Salmivalli, 2012), most existing studies focused on the individual characteristics of bullies and victims while failing to fully appreciate its nature as a group phenomenon. At the same time, growing evidence highlights the significance of peer relationship and social network factors in students’ involvement in bullying. Some researchers regard bullying as a group process and suggest that more effort should be devoted to uncover the peer dynamics that contribute to the student’s role in such incidents (Salmivalli, 2010). In recent years, the researchers have begun to analyze the issue of school violence from the perspective of peer clusters and social networks (e.g., Swartz, Reyns, Wilcox, & Dunham, 2012).

52

Wei and Lee

Despite the increasing recognition of the importance of social network dynamics in b­ ullying, only scattered investigations so far have used this approach, and many of them share several common shortcomings. For example, when relational factors such as peer relationships and popularity are included in analysis, they are usually measured by self-reports and suffer considerable subjective biases. Responses to items such as “I feel I have many friends in school” reflect merely personal perception and offer little information on the specific friendship matrices. In addition, past research has relied largely on cross-sectional data, which reveal only associations but make it difficult, if not impossible, to make causal inferences. It has been long suggested that longitudinal data is preferable for clarifying the direction of causality and enhancing the predictive power of the findings. Furthermore, the peer network is constantly changing, and many traditional analytic techniques are unable to capture its dynamic features.

THE PRESENT STUDY In response to the previously mentioned issues, this study proposes an alternative approach by using social network analysis to examine students’ involvement in physical bullying. Social relationships and group structures heavily influence the individual’s action. For adolescents, peer group is the primary interpersonal context that shapes and maintains their behavior. The concept of social network analysis has recently been applied to several adolescent health topics, including young people’s eating behavior (Fletcher, Bonell, & Sorhaindo, 2011) and smoking behavior (Seo & Huang, 2012). In the middle school system of Taiwan, as in many other countries, every student is assigned to one homeroom class and stays with a fixed group of classmates every day throughout their school years. As a result, bullying in Taiwan occurs primarily between students within the homeroom class (Wei, Jonson-Reid, & Tsao, 2007). The class thus constitutes a bounded group that is suitable for social network analysis, and this study examines students’ involvement in bullying by their position in the class-wide peer network rather than by their self-reported scores. Moreover, to explore changes in the individual’s network position over time, a longitudinal dataset with multiple waves of social network measures is used, and the potential interaction between bully/victim position and friendship in predicting further bullying involvement is explored. It is hoped that by adopting a social network perspective, this study not only addresses several limitations of past research but also demonstrates the use of network in predicting student’s physical bullying involvement. Specifically, three research questions are explored in this study:



1. How do individual-level factors, including gender, self-esteem, academic performance, parental marital status, paternal education, and family economic status, predict student’s physical bully and victim centrality in the class-wide peer network? 2. How do network factors, such as students’ physical bully and victim centrality at time t and centrality in friendship network, predict their bully and victim centrality at t 1 1? 3. How do previous physical bully and victim centrality interact with students’ friendship network in predicting students’ subsequent involvement in bullying?

METHODS Participants This study followed four 7th-grade classes from two middle schools in Taipei, Taiwan for 1 entire year. The average class size was 31.25, with the largest class size being 34

Individual and Social Network Predictors of Physical Bullying 53

and the smallest, 27. There were 125 students recruited, and five waves of data were collected during the period. Overall, 53.6% of the students were boys and 46.4% were girls. Furthermore, 86.2% of the respondents reported that their parents were married and living together, whereas about 9% of the parents were divorced.

Instruments Physical Bully and Victim Centrality. As the key dependent variables, every student’s bully and victim centrality in the class was assessed five times during the year. A class roster with all the students’ names was distributed to each respondent. A behavioral description of physical bullying (hitting, kicking, pushing, etc.) was also provided. Students were asked to indicate the classmates that often bully them physically. Using the information, we constructed a binary matrix of bullying network, B, where element bij denoted whether student i nominated student j as the one who ever physically bullied i in the time period. The number of nominations that a student received from all classmates was used to calculate his or her physical bully degree centrality (Freeman, 1978). Likewise, a student’s victim centrality was computed based on the number of ties that the individual directed to others (i.e., the number of bullies that the student nominated). Bully centrality for a class a at t is a vector of column sum of Bat, whereas victim centrality is a vector of row sum of Bat. Popularity and Embeddedness in Friendship Network. A class roster with all students’ names was distributed to each respondent. Students were asked to nominate their friends from the list. From this, we constructed a binary matrix of friendship network, F, where element fij denoted whether student i nominated student j as his or her friend at t. The number of nominations that each student received was used to calculate his or her popularity in a at t, which was a vector of column sum of Fat. Embeddedness in friendship networks for each student was calculated using a clustering coefficient (Watts & Strogatz, 1998), which measures how tightly the actor is connected within his or her friendship network. When two of his or her friends are also friends with each other, the student has a triangular relationship. In theory, if (s)he has friends n, the maximum number of triangles (s)he can have is n (n 2 1)/2. The clustering coefficient for a student is the actual number of triangles divided by the maximum number. We symmetrized F to obtain clustering coefficients for individual students (see Reagans & McEvily, 2003; Sosa, 2011). When the clustering coefficient is high, the student’s friendship ties are considered strong because the two dyadic relationships are bound by the extra relationships among his or her friends (Granovetter, 1973). Likewise, the student is tightly embedded in the friendship network (s)he belongs to because a dyadic friendship is monitored and facilitated by other friends who are also friends to both parties (Gulati & Westphal, 1999). To illustrate the four network measures calculated from bullying and friendship networks, consider the network diagrams depicted in Figure 1 for Class 1 at Time 1. The first diagram indicates who bullied whom and how much in Class 1 at Time 1. Larger node size indicates that the student got a larger number of nominations in the bullying survey. On the other hand, nodes in the friendship network in Class 1 at Time 1 was weighted by the number of friend nomination and clustering coefficients, which made the second and third diagrams.

54

Centrality in physical bullying network (Class 1, Time 1)

Popularity in friendship network (Class 1, Time 1)

Embeddedness in friendship network (Class 1, Time 1)

Figure 1.  Centralities in bullying network, popularity in friendship network, and embeddedness in friendship network (Class 1, Time 1).

Wei and Lee

Individual and Social Network Predictors of Physical Bullying 55

Individual Factors. Several individual variables identified in the literature were assessed through self-report surveys, which include the student’s gender (0 5 female, 1 5 male), self-esteem (measured by the 10-item Rosenberg Self-Esteem Scale [Rosenberg, 1965]), and his or her midterm exam score rank in the first semester as an indicator of academic performance. The score rank was based on a total exam score across all courses, and a higher percentile rank represents better academic performance. Family Condition. Information on several relevant family socioeconomic factors was also collected through students’ self-reports. The respondents were asked to indicate their parents’ marital status on a 7-point scale ranging from “married and living together” to “both parents passed away.” Because most parents in those families were married and living together, the responses were further dichotomized to create a binary variable (1 5 ­married and living together, 0 5 else). Students also provided information on their parents’ educational degree (1 5 illiterate to 6 5 graduate school), and the educational levels of the two parents were averaged into an indicator of parental education level. Finally, the economic status of the family was assessed on a 5-item scale (5 5 rich to 1 5 very poor).

Procedure Invitation letters were sent to local middle schools in Taipei. Two public schools agreed to participate in the study. From each school, two seventh-grade classes were recruited as the sample for this study. Seventh-grade students were selected because it was their first semester in middle school, and the homeroom classes were just assembled with unfamiliar members, providing a great opportunity to observe their group dynamics from the beginning. Written consents were obtained, and research descriptions were distributed to the students and teachers. All students in the four classes agreed to participate in the survey. Social network data was collected at five time points during the school year to record the physical bully/victim networks along with the friendship matrix within the class. The questionnaire battery requires about 20 min to complete and was administered during the lunch break or an independent study session; participants returned the questionnaires in sealed envelopes. The students also provided individual and family information through self-reported surveys early in the year.

Data Analysis The focus of this study was to predict students’ longitudinal physical bullying involvement by examining individual-level and network-level factors. Because the data contains five waves of assessment from four classrooms in two schools, mixed model analysis with both fixed effect and random effect was used as the major statistical strategy of this study. These models are particularly suitable when repeated measures and clustered data are involved. In addition, the dependent variable and some of the independent variables in this study were network factors, which made the models become mixed endogenous–exogenous. In this case, maximum likelihood yields more efficient estimation (Doreian, 1981). Hence, the models were fitted using STATA 12.0 MP’s xtmixed module with maximum likelihood estimation. Students were assumed to be nested within their classes, and multilevel regressions were conducted to allow for the inclusion of random effects at student and classroom levels. Time-constant, subject-specific variables as well as the subjects’ friendship and bully/victim network centrality at t were employed to predict individual change in bullying and victimization (i.e., bully/victim network centrality at t 1 1). At the same time, fixed effects of time and classrooms were included to control for temporal and classroomspecific confounding influences (four classrooms and five waves of data).

56

Wei and Lee

RESULTS Table 1 summarizes the results of the models predicting students’ centrality in victim/ bully networks from t to t 1 1 (approximately 1 month apart) during the five waves. The influences of individual level characteristics exhibited the same direction for both centralities. Although boys (gender 5 1) were more actively involved in being either a victim or a bully over time, parents’ educational level was negatively associated with both centralities. When parents are neither divorced nor separated, the students are less likely to be a bully. Apparently, higher self-esteem correlates negatively with being a victim. Interestingly, higher family economic status correlated positively with the risk of being a victim. The effects of network factors suggest two interesting points about school violence. First, the growth rate in centrality in either the victim or bully network is more than 1 (1.25 for victim, 1.54 for bully). It means that once classified as a victim or a bully, the student increasingly moves toward the center of the web of violence. Second, popularity and embeddedness in friendship network do not appear to be directly related to either being a victim or a bully. However, the interactions between the network factors and victim/bully centralities are significant. It suggests that the effects of the network factors vary with the level of students’ past involvement in victimization and bullying (e.g., Faunce, 1984). According to the coefficients in Table 1, the effect of embeddedness on the victim centrality at t 1 1 is 1.15, 22.53* victim centrality at t. Likewise, the embeddedness effects on bully centrality at t 1 1 is 20.19, 20.65* bully centrality at t, and the effects of popularity on victim and bully centrality at t 1 1 are 20.05, 10.07* victim centrality at t and 2.001, 2.05*bully centrality at t, respectively.

Figure 2 plots the relations between the effects of the network factors and victim/bully centralities. In the model predicting victim centrality at t 1 1, embeddedness appears to be decreasing the victim centrality at t 1 1 unless the student has low score on victim centrality at t. Embeddedness also decreases the bully centrality at t 1 1. It is consistent with sociological studies of school, which argues that “fictive kinship” of friends offers psychological and social supports to students (Carter, 2007). On the other hand, popularity exhibits inconsistent signs. Although popularity largely decreases the bully centrality at t 1 1, it turns out to be increasing the victim centrality at t 1 1. Being popular is generally associated with successful school life for students (see Hansell, 1985). However, popularity can also generate grievance from schoolmates because the popular student has difficulty in satisfying all the social demands placed on them (Eder, 1985). The lower left plot in Figure 2 indicates that popularity is more likely to put the student in harm’s way when his or her victim centrality is higher at t.

DISCUSSION Students are constantly interacting with peers in school, and these interactions and relationships constitute social networks that have a substantial effect on several adolescent behaviors, including their involvement in school bullying as bullies and victims. Past

Individual and Social Network Predictors of Physical Bullying 57

TABLE 1.  Multilevel Mixed Effects Regressions Predicting Centrality in Victim/ Bully Network Centrality in Victim Network (t 1 1) Centrality in victim network (t)

1.2500 (0.3900)***

Centrality in Bully Network (t 1 1) 20.0700 (0.0200)***

Centrality in bully network (t)

20.0200 (0.0800)

Popularity (t)

20.0500 (0.0300)

20.0020 (0.0140)

Embeddedness in friendship network (t)

1.1500 (0.8600)

20.1900 (0.4300)

Centrality in victim network (t) 3 popularity (t)

0.0700 (0.0100)***

Centrality in bully network (t) 3 popularity (t) Centrality in victim network (t) 3 embeddedness in friendship network (t)

— 22.5300 (0.6000)***

Centrality in bully network (t) 3 embeddedness in friendship network (t)

—

1.5400 (0.1900)***

— 20.0500 (0.0100)*** —

20.6500 (0.2000)***

Family economic status

0.9000 (0.2300)***

0.1041 (0.0900)

Gender

0.9200 (0.2500)***

0.2500 (0.1000)***

Parents’ marital status

20.5000 (0.3900)

20.3900 (0.1500)***

Parents’ education

20.2000 (0.0800)***

20.0700 (0.0300)**

Self-esteem

20.0700 (0.0200)***

20.0100 (0.0100)

Academic performance

20.0200 (0.0100)

20.0030 (0.0100)

Class 1

0.3700 (0.3300)

0.1200 (0.1300)

Class 2

0.4400 (0.3200)

0.3300 (0.1200)***

Class 3

20.2700 (0.3200)

0.0200 (0.1200)

Wave 1

0.3700 (0.2600)

0.2100 (0.1300)

Wave 2

20.0300 (0.2500)

20.2000 (0.1200)

Wave 3

20.1400 (0.2300)

20.1000 (0.1200)

1.2800 (1.3900)

1.2000 (0.6100)

404

404

Constant The number of observations Wald

x2(df)

138.9000 (18.0000)

Note. Standard error is in parenthesis. **p , .05. ***p , .001.

797.5000 (18.0000)

8

6

4

2

50 40 30 20 10 0

0

Wei and Lee

0

10

20

30

0

10

20

30

5

10

0

5

10

1

0

0

2

15

3

10

4

5 0

Effect of popularity

20

Effect of embeddedness

58

Victim centrality

Bully centrality

Figure 2.  Relations between effects of popularity/embeddedness and victim/bully centralities.

research has identified several individual-level risk factors, and some of these factors were examined empirically in this study with a sample of Taiwanese middle school students. Consistent with prior findings (e.g., Scheithauer et al., 2006; Wang et al., 2009), this analysis found that boys are more likely to be both physical bullies and victims. This is consistent with previous research showing that male adolescents have a higher engagement in direct aggression compared to female adolescents (for a review, see Card, Stucky, Sawalani, & Little, 2008). Similarly, low self-esteem was found to be predictive of later victimization. Although academic performance exhibited a negative correlation with future victimization and bullying, the effect did not reach the significant level. Interestingly, one recent study on Taiwanese adolescents revealed the moderating effect of Machiavellianism on the relationship between bullying and school adjustment (Wei & Chen, 2012). For those who were low in Machiavellianism condition, bullying was negatively linked to academic performance, whereas no significant association was found for the high-Machiavellianism group. Such findings highlighted the heterogeneity of bullies and suggest that certain aggressive adolescents do not necessarily perform poorly at school. As for family factors, lower parental education level predicts a youth’s likelihood of being a physical bully and victim. Having parents married and living together also reduced the risk of being a bully over time, which largely corresponds to the previous findings in the school bullying literature. However, higher family economic status was associated with a higher risk of being victimized. One possible explanation for this unexpected finding is that the victim status in this study is determined by the respondent’s nominations of classmates who physically bullied him or her, reflecting the nominator’s subjective

Individual and Social Network Predictors of Physical Bullying 59

perception. Past research has suggested that people from different socioeconomic backgrounds hold different attitudes toward violence (Heimer, 1997), which may result in higher identification of peers’ physical acts (e.g., pushing, shoving) as aggressive bullying among ­adolescents from more affluent families because such acts are less tolerated in their culture. Besides evaluating the effects of individual-level variables, the results of this study highlight the importance of peer network factors in students’ physical bullying involvement. First, a student’s centrality in the victim network at t predicts the student’s victim centrality at t 1 1 positively. Although it is insignificant, the sign of the bully centrality’s effect on subsequent victimization is negative. Similarly, the student’s bully and victim centrality at t were both found to be positive and negative predictors of further bullying perpetration. Such results not only suggest a nonoverlap between bullies and victims but also reveal the continuity and even escalation of bullying network positions in the classroom. In addition to the stability of individual attributes, peers are also likely to contribute to this chronic aggression and victimization by conferring reputations on the offender and victim that secure them into fixed roles (DeRosier, Cillessen, Coie, & Dodge, 1994). In other words, the reputations of the bully and victim can shape the ways in which peers interact with them and lock the two parties into relatively stable positions in the social network (Biddle, 1986; Thye, 2000). Regarding the peer friendship network, popularity and embeddedness in friendship had no main effect on bully centrality, showing that physical bullies are not necessarily rejected by peers or do not necessarily lack friends. In fact, recent studies showed that bullies can manage to avoid negative social consequences of their behavior through strategic aggression (Veenstra, Lindenberg, Munniksma, & Dijkstra, 2010; Wei & Chen, 2012). Interestingly, popularity and embeddedness in friendship network did not significantly contribute to lower risk of victimization. To have a better understanding of these results, the interaction between peer friendship and previous bully/victim status in predicting subsequent bully/victim centrality was further explored. It was found that popularity and embeddedness in friendship largely reduced the risk of bullying, especially for those with higher previous engagement in bullying. This finding is consistent with longitudinal research by Kreager (2004), who showed that social isolation in conjunction with problematic peer interaction in school significantly increases later delinquent behavior. These findings suggest that embeddedness in friendship can buffer such a trend, and it may be used as a potential social network mechanism for decreasing bullying behavior and curving negative role taking (Loch, Huberman, & Stout, 2000). At the same time, a close examination of the positive interaction effect of popularity and prior victim status in p­ redicting further victimization reveals that friendship acts as a protective factor only for the students with low previous victimization. For those who had been highly victimized in the past, popularity actually increases their future risk of being bullied. A discrepancy in the definition of relationship nature may occur because a student’s friendship was calculated based on the number of classmates who nominated the target person as a friend, whereas the person’s victim status was assessed by his or her own nominations of classmates who physically bullied him/her. It is possible that one student’s gesture of playful fighting may have been perceived as harassment by another. Moreover, in one study of Taiwanese adolescents, Wei and Jonson-Reid (2011) showed that friendship and victimization might coexist for some students because some students reported that their friends bullied them. Further investigation needs to clarify the problematic interaction patterns in these companionable yet harmful relationships.

60

Wei and Lee

Limitations Despite the strengths in data quality and analytic approach discussed earlier, this study has several limitations that are worth noting. First, past research has accumulated a large ­number of individual risk factors of school bullying, such as one’s delinquency and probullying beliefs, many of which are not included in this study. In addition, nonphysical types of bullying, such as rumor spreading and social isolation, are highly relevant to social network dynamics, but they are not examined here. Second, the interplay between peer relationship and bullying involvement revealed in this study suggests the potential existence of complicated social interactions among adolescents that awaits future exploration. For example, it was found that bullies often manage to minimize social cost by choosing victims who are not likely to be defended by same-gender peers (e.g., victims of male bullies are rejected by males only; Veenstra et al., 2010). More research effort can be paid to further clarify the mechanisms that contribute to the development and maintenance of the bully/victim structure within classroom. In addition, the bullying networks were constructed from the victims’ nominations of the students who bullied them. Such measures reflect the nominators’ self-report and may be biased by their personal perception. Multiple sources of data, such as teacher ratings or classroom observations, can be used in future research to provide a more comprehensive and balanced picture of the peer network. Finally, the four-classroom sample size of this study is relatively limited. However, with multiple waves of follow-up data collection, the actual number of observations was comparatively large, helping to achieve statistical stability and adding to the robustness of results.

CONCLUSIONS This study is the first to examine physical bullying involvement among Taiwanese middle school students from a social network perspective. It is shown that besides individual and demographic characteristics, network factors also have substantial effects on students’ involvement in physical bullying. Multiple waves of social network data further allow a better appreciation of the relational dynamics in peer groups that contribute to one’s involvement in bullying as perpetrators and as victims. The interaction effect between friendship and bullying is of particular interest. In short, popularity in peer groups is not always rewarding. Because peer classroom contexts heavily influence the behavior and adjustment of adolescents, school health professionals are challenged with the task to design and implement group-level strategies to enhance students’ development and well-being. This study highlights the significance of social network factors in predicting adolescents’ physical bullying involvement and provides school health professionals with an alternative framework to understand this common aggression among youth. Teachers, school counselors, psychologists, and social workers should address student bullying as a group process phenomenon. Traditional focus on individual risks and vulnerabilities may misguide attention and preclude the possibility to enhance classroom climate and peer dynamics. In contrast, it is argued that school bullying is essentially a peer interaction issue, and more effort should be devoted to monitor interpersonal relationships and adjust network structures. By identifying the within-class network factors operating in school bullying, this study hopes to stimulate further research and service programs to target group dynamics as potentially effective measures to prevent physical bullying and promote students’ health in school.

Individual and Social Network Predictors of Physical Bullying 61

REFERENCES Andreou, E. (2000). Bully/victim problems and their association with psychological constructs in 8- to 12-year-old Greek schoolchildren. Aggressive Behavior, 26(1), 49–56. http://dx.doi .org/10.1002/(SICI)1098-2337(2000)26:1{49::AID-AB4}3.0.CO;2-M Bender, D., & Lösel, F. (2011). Bullying at school as a predictor of delinquency, violence and other anti-social behaviour in adulthood. Criminal Behaviour and Mental Health, 21(2), 99–106. http://dx.doi.org/10.1002/cbm.799 Beran, T. N., Hughes, G., & Lupart, J. (2008). A model of achievement and bullying: Analyses of the Canadian National Longitudinal Survey of Children and Youth data. Educational Research, 50(1), 25–39. http://dx.doi.org/10.1080/00131880801920379 Biddle, B. J. (1986). Recent development in role theory. Annual Review of Sociology, 12, 67–92. Boulton, M. J., Chau, C., Whitehand, C., Amataya, K., & Murray, L. (2009). Concurrent and shortterm longitudinal associations between peer victimization and school and recess liking during middle childhood. British Journal of Educational Psychology, 79(2), 207–221. http://dx.doi .org/10.1348/000709908X336131 Card, N. A., Stucky, B. D., Sawalani, G. M., & Little, T. D. (2008). Direct and indirect aggression during childhood and adolescence: A meta-analytic review of gender differences, intercorrelations, and relations to maladjustment. Child Development, 79(5), 1185–1229. http://dx.doi .org/10.1111/j.1467-8624.2008.01184.x Carter, D. J. (2007). Why the black kids sit together at the stairs: The role of identity-affirming counter-spaces in a predominantly white high school. The Journal of Negro Education, 76(4), 542–554. Craig, W., Harel-Fisch, Y., Fogel-Grinvald, H., Dostaler, S., Hetland, J., Simons-Morton, B., . . . Pickett W. (2009). A cross-national profile of bullying and victimization among adolescents in 40 countries. International Journal of Public Health, 54, 216–224. de Bruyn, E. H., Cillessen, A. N., & Wissink, I. B. (2010). Associations of peer acceptance and perceived popularity with bullying and victimization in early adolescence. The Journal of Early Adolescence, 30(4), 543–566. http://dx.doi.org/10.1177/0272431609340517 DeRosier, M. E., Cillessen, A. N., Coie, J. D., & Dodge, K. A. (1994). Group social context and children’s aggressive behavior. Child Development, 65(4), 1068–1079. http://dx.doi .org/10.2307/1131305 Doreian, P. D. (1981). Estimating linear models with spatially distributed data. Sociological Methodology, 12, 359–388. Due, P., Merlo, J., Harel-Fisch, Y., Damsgaard, M., Holstein, B. E., Hetland, J., . . . Lynch, J. (2009). Socioeconomic inequality in exposure to bullying during adolescence: A comparative, crosssectional, multilevel study in 35 countries. American Journal of Public Health, 99(5), 907–914. http://dx.doi.org/10.2105/AJPH.2008.139303 Eder, D. (1985). The cycle of popularity: Interpersonal relations among female adolescents. Sociology of Education, 58(3), 154–165. Farrington, D. P., Loeber, R., Stallings, R., & Ttofi, M. M. (2011). Bullying perpetration and victimization as predictors of delinquency and depression in the Pittsburgh Youth Study. Journal of Aggression, Conflict and Peace Research, 3(2), 74–81. Faunce, W. A. (1984). School achievement, social status, and self-esteem. Social Psychology Quarterly, 47(1), 3–14. Fekkes, M., Pijpers, F. M., Fredriks, A., Vogels, T., & Verloove-Vanhorick, S. (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(5), 1568–1574. http://dx.doi .org/10.1542/peds.2005-0187 Fisher, H. L., Moffitt, T. E., Houts, R. M., Belsky, D. W., Arseneault, L., & Caspi, A. (2012). Bullying victimisation and risk of self harm in early adolescence: Longitudinal cohort study. BMJ: British Medical Journal, 344(7855), 1–9.

62

Wei and Lee

Fletcher, A., Bonell, C., & Sorhaindo, A. (2011). You are what your friends eat: Systematic review of social network analyses of young people’s eating behaviours and bodyweight. Journal of Epidemiology and Community Health, 65(6), 548–555. http://dx.doi.org/10.1136/ jech.2010.113936 Freeman, L. C. (1978). Centrality in social networks: Conceptual clarification. Social Networks, 1, 215–239. Gladstone, G. L., Parker, G. B., & Malhi, G. S. (2006). Do bullied children become anxious and depressed adults? A cross-sectional investigation of the correlates of bullying and anxious depression. Journal of Nervous and Mental Disease, 194(3), 201–208. http://dx.doi .org/10.1097/01.nmd.0000202491.99719.c3 Granovetter, M. S. (1973). The strength of weak ties. American Journal of Sociology, 78(6), 1360–1380. Gulati, R., & Westphal, J. D. (1999). Cooperative or controlling? The effects of CEO-board relations and the content of interlocks on the formation of joint ventures. Administrative Science Quarterly, 44(3), 473–506. Hansell, S. (1985). Adolescent friendship networks and distress in school. Social Forces, 63(3), 698–715. Heimer, K. (1997). Socioeconomic status, subcultural definitions, and violent delinquency. Social Forces, 75, 799–833. Hemphill, S. A., Kotevski, A., Tollit, M., Smith, R., Herrenkohl, T. I., Toumbourou, J. W., & Catalano, R. F. (2012). Longitudinal predictors of cyber and traditional bullying perpetration in Australian secondary school students. Journal of Adolescent Health, 51(1), 59–65. http:// dx.doi.org/10.1016/j.jadohealth.2011.11.019 Huitsing, G., Veenstra, R., Sainio, M., & Salmivalli, C. (2012). “It must be me” or “It could be them?”: The impact of the social network position of bullies and victims on victims’ adjustment. Social Networks, 34(4), 349–386. http://dx/doi.org/10.1016/j.socnet.2010.07.002 Jenson, J. M., Dieterich, W. A., Brisson, D., Bender, K. A., & Powell, A. (2010). Preventing childhood bullying: Findings and lessons from the Denver Public Schools Trial. Research on Social Work Practice, 20, 509–517. Juvonen, J., Wang, Y., & Espinoza, G. (2011). Bullying experiences and compromised academic performance across middle school grades. Journal of Early Adolescence, 31(1), 152–173. http:// dx.doi.org/10.1177/0272431610379415 Khoury-Kassabri, M., Benbenishty, R., & Astor, R. (2005). The effects of school climate, socioeconomics, and cultural factors on student victimization in Israel. Social Work Research, 29(3), 165–180. http://dx.doi.org/10.1093/swr/29.3.165 Kim, Y., Boyce, W., Koh, Y., & Leventhal, B. L. (2009). Time trends, trajectories, and demographic predictors of bullying: A prospective study in Korean adolescents. Journal of Adolescent Health, 45(4), 360–367. http://dx.doi.org/10.1016/j.jadohealth.2009.02.005 Klomek, A., Marrocco, F., Kleinman, M., Schonfeld, I. S., & Gould, M. S. (2007). Bullying, depression, and suicidality in adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 46(1), 40–49. http://dx.doi.org/10.1097/01.chi.0000242237.84925.18 Kreager, D. (2004). Strangers in the halls: Isolation and delinquency in school networks. Social Forces, 83(1), 351–390. http://dx.doi.org/10.1353/sof.2004.0117 Lauritsen, J. L., & Heimer, K. (2008). The gender gap in violent victimization, 1973-2004. Journal of Quantitative Criminology, 24(2), 125–147. http://dx.doi.org/10.1007/s10940-008-9041-y Lauritsen, J. L., Heimer, K., & Lynch, J. P. (2009). Trends in the gender gap in violent offending: New evidence from the National Crime Victimization Survey. Criminology: An Interdisciplinary Journal, 47(2), 361–399. http://dx.doi.org/10.1111/j.1745-9125.2009.00149.x Loch, C. H., Huberman, B. A., & Stout, S. (2000). Status competition and performance in work groups. Journal of Economic Behavior & Organization, 43, 35–55. Mouttapa, M., Valente, T., Gallaher, P., Rohrbach, L., & Unger, J. B. (2004). Social network ­predictors of bullying and victimization. Adolescence, 39(154), 315–335.

Individual and Social Network Predictors of Physical Bullying 63 O’Moore, M. M., & Kirkham, C. C. (2001). Self-esteem and its relationship to bullying behaviour. Aggressive Behavior, 27(4), 269–283. http://dx.doi.org/10.1002/ab.1010 Pellegrini, A. D. (2002). Bullying, victimization, and sexual harassment during the transition to middle school. Educational Psychologist, 37(3), 151–163. http://dx.doi.org/10.1207/ S15326985EP3703_2 Pellegrini, A. D., Bartini, M., & Brooks, F. (1999). School bullies, victims, and aggressive victims: Factors relating to group affiliation and victimization in early adolescence. Journal of Educational Psychology, 91(2), 216–224. http://dx.doi.org/10.1037/0022-0663.91.2.216 Radliff, K. M., Wheaton, J. E., Robinson, K., & Morris, J. (2012). Illuminating the relationship between bullying and substance use among middle and high school youth. Addictive Behaviors, 37(4), 569–572. http://dx.doi.org/10.1016/j.addbeh.2012.01.001 Reagans, R., & McEvily, B. (2003). Network structure and knowledge transfer: The effects of cohesion and range. Administrative Science Quarterly, 48(2), 240–267. Reijntjes, A., Vermande, M., Goossens, F. A., Olthof, T., van de Schoot, R., Aleva, L., & van der Meulen, M. (2013). Developmental trajectories of bullying and social dominance in youth. Child Abuse & Neglect, 37(4), 224–234. http://dx.doi.org/10.1016/j.chiabu.2012.12.004 Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press. Salmivalli, C. (2010). Bullying and the peer group: A review. Aggression and Violent Behavior, 15(2), 112–120. http://dx.doi.org/10.1016/j.avb.2009.08.007 Scheithauer, H., Hayer, T., Petermann, F., & Jugert, G. (2006). Physical, verbal, and relational forms of bullying among German students: Age trends, gender differences, and correlates. Aggressive Behavior, 32(3), 261–275. http://dx.doi.org/10.1002/ab.20128 Sentse, M., Scholte, R., Salmivalli, C., & Voeten, M. (2007). Person-group dissimilarity in involvement in bullying and its relation with social status. Journal of Abnormal Child Psychology, 35(6), 1009–1019. http://dx.doi.org/10.1007/s10802-007-9150-3 Seo, D. C., & Huang, Y. (2012). Systematic review of social network analysis in adolescent cigarette smoking behavior. Journal of School Health, 82(1), 21–27. Sosa, M. E. (2011). Where do creative interactions come from? The role of tie content and social networks. Organization Science, 22(1), 1–21. Swartz, K., Reyns, B. W., Wilcox, P., & Dunham, J. R. (2012). Patterns of victimization between and within peer clusters in a high school social network. Violence and Victims, 27(5), 710–729. Thornberg, R. (2011). ‘She’s weird!’—The social construction of bullying in school: A review of qualitative research. Children & Society, 25, 258–267. http://dx.doi.org/10.1111/j .1099-0860.2011.00374.x Thye, S. R. (2000). A status value theory of power in exchange relations. American Sociological Review, 65(3), 407–432. Tom, S. R., Schwartz, D., Chang, L., Farver, J. M., & Xu, Y. (2010). Correlates of victimization in Hong Kong children’s peer groups. Journal of Applied Developmental Psychology, 31(1), 27–37. http://dx.doi.org/10.1016/j.appdev.2009.06.002 Ttofi, M. M., Farrington, D. P., Lösel, F., & Loeber, R. (2011). The predictive efficiency of school bullying versus later offending: A systematic/meta-analytic review of longitudinal studies. Criminal Behaviour and Mental Health, 21(2), 80–89. http://dx.doi.org/10.1002/cbm.808 Undheim, A., & Sund, A. (2010). Prevalence of bullying and aggressive behavior and their relationship to mental health problems among 12- to 15-year-old Norwegian adolescents. European Child & Adolescent Psychiatry, 19(11), 803–811. http://dx.doi.org/10.1007/s00787-010-0131-7 Veenstra, R., Lindenberg, S., Munniksma, A., & Dijkstra, J. (2010). The complex relation between bullying, victimization, acceptance, and rejection: Giving special attention to status, affection, and sex differences. Child Development, 81(2), 480–486. http://dx.doi.org/10.1111/j .1467-8624.2009.01411.x Von Marées, N., & Petermann, F. (2010). Bullying in German primary schools: Gender differences, age trends and influence of parents’ migration and educational backgrounds. School Psychology International, 31(2), 178–198. http://dx.doi.org/10.1177/0143034309352416

64

Wei and Lee

Wang, J., Iannotti, R. J., & Nansel, T. R. (2009). School bullying among adolescents in the United States: Physical, verbal, relational, and cyber. Journal of Adolescent Health, 45(4), 368–375. http://dx.doi.org/10.1016/j.jadohealth.2009.03.021 Watts, D. J., & Strogatz, S. H. (1998). Collective dynamics of ‘small-world’ networks. Nature, 393, 440–442. Wei, H., & Chen, J. (2009). Social withdrawal, peer rejection, and peer victimization in Taiwanese middle school students. Journal of School Violence, 8(1), 18–28. http://dx.doi .org/10.1080/15388220802067755 Wei, H. S., & Chen, J. K. (2012). The moderating effect of Machiavellianism on the relationships between bullying, peer acceptance, and school adjustment in adolescents. School Psychology International, 33, 345–363. Wei, H. S., & Jonson-Reid, M. (2011). Friends can hurt you: Examining the coexistence of friendship and bullying among early adolescents. School Psychology International, 32(3), 244–262. Wei, H. S., Jonson-Reid, M., & Tsao, H. L. (2007). Bullying and victimization among Taiwanese 7th graders: A multi-method assessment. School Psychology International, 28(4), 479–500. Wei, H. S., Sun, S. W., Liu, S. F., & Chen, J. K. (2011). 2010 Report on the life condition of children and adolescents in Taiwan. Taichung, Taiwan: Child Bureau. Wei, H., & Williams, J. (2004). Relationship between peer victimization and school adjustment in sixth-grade students: Investigating mediation effects. Violence and Victims, 19(5), 557–571. http://dx.doi.org/10.1891/vivi.19.5.557.63683 Acknowledgments. This research is funded by the MSIP (Ministry of Science, ICT & Future Planning), Korea in the ICT R & D Program 2014. Correspondence regarding this article should be directed to Hsi-Sheng Wei, PhD, National Taipei University, Department of Social Work, 151, University Rd., San Shia, New Taipei City 237, Taiwan. E-mail: [email protected]

Violence and Victims, Volume 29, Number 2, 2014

Psychometric Properties of the Cyberbullying Questionnaire (CBQ) Among Mexican Adolescents Manuel Gámez-Guadix, PhD University of Deusto, Vizcaya, Spain

Fabiola Villa-George, PhD National Autonomous University of México, México

Esther Calvete, PhD University of Deusto, Vizcaya, Spain The first objective of this study was to analyze the psychometric properties of the Cyberbullying Questionnaire (CBQ), an instrument for measuring the perpetration and victimization of bullying via new technologies for adolescents. The second objective was to analyze gender differences in the prevalence of cyberbullying. The study sample consisted of 1,491 Mexican adolescents (52.4% male and 47.6% female) with a mean age of 14.51 years (SD 5 1.57, range 5 12–18). A confirmatory factor analysis of the CBQ indicated a good fit of a model consisting of two factors designated as “perpetration” and “victimization.” The internal consistencies for these subscales were adequate. Furthermore, multiple-group-covariance-structure analysis with the Mexican and a Spanish sample (N 5 1008; 55.7% girls; mean age 5 15.23 years, SD 5 1.4) indicated equivalence of the factor structure of the CBQ across samples. An analysis of the relationship between the CBQ and other variables—such as the justification of cyberbullying, impulsivity, and depression—provided additional data supporting the construct validity of the instrument. Regarding gender differences in the prevalence of CB, perpetration was significantly higher for males than for females, whereas no differences were found for victimization. Finally, we discuss the contributions of this work to the field of study.

Keywords: cyberbullying; CBQ; internet bullying; psychometric properties; justification of violence

C

yberbullying (CB), the harassment of others via new technologies, is a growing phenomenon with important consequences for its victims. CB is defined as a series of recurring, intentionally aggressive behaviors that is carried out by a group or individual using electronic means. The aim of such behavior is to threaten, embarrass, or intimidate a victim who cannot easily defend himself or herself (Patchin & Hinduja, 2006; Smith, Mahdavi, Carvalho, & Tippett, 2006). This aggressive behavior is implemented by using cell phones, e-mail, online chats, and/or online spaces such as Facebook, Messenger, or personal blogs (Calvete, Orue, Estévez, Villardón, & Padilla, 2010; Li, 2008).

© 2014 Springer Publishing Company

65

66

Gámez-Guadix et al.

The behaviors that are relevant to CB include sending threatening or insulting e-mails, sending or uploading images or starting rumors that are cruel or harmful to the victim’s reputation, or “happy slapping,” in which recorded images of a person who is attacked or humiliated are taken and disseminated (Kiriakidis & Kavoura, 2010; Smith et al., 2006; Willard, 2006). Despite the growing interest in these new forms of violence and the significant impacts they can have on their victims, there are only a few examples of studies on measurement issues in this area (Calvete et al., 2010; Slonje, Smith, & Frisén, 2013). Incidents of CB are common among adolescents. In a recent review, Tokunaga (2010) notes that most studies find that approximately 20%–40% of young people have been victims of CB. However, these rates vary widely across studies (Estévez, Villardón, Calvete, Padilla, & Orue, 2010; Kowalski & Limber, 2007; Ybarra & Mitchell, 2004). Regarding CB perpetration, studies indicate that between 11% and 44% of adolescents have been cyberbullies (Calvete et al., 2010; Kowalski & Limber, 2007). The common methods of CB in these studies included instant messaging, chat rooms, and e-mail. These wide variations among studies may be caused, among other factors, by the rapid development of new technologies (e.g., the generalization of smartphones), the time interval considered (e.g., “ever,” “the last 6 months,” etc.), and the range of different types of CB considered in each study. In developing countries such as Mexico, CB research is scarce. According to data from the World Internet Project (WIP, 2013) in Mexico, more than 38.4% of Internet users were 18 years old or younger in 2012. Furthermore, the number of users is increasing substantially. For instance, in 2012, the number of Internet users younger than 18 years increased by 25.6% with respect to the 2011 data (WIP, 2013). These data suggest that the concerns related to problematic use of the Internet also may be increasing significantly. Nonetheless, empirical data on the prevalence of CB in Mexico is virtually inexistent. In one of the few available studies, Lucio (2009) found in a sample of 1,066 Mexican adolescents that the most frequent forms of CB were threatening over the Internet (perpetration: 22.2%; victimization: 16.9%), sexually harassing over the Internet (perpetration: 10.0%; victimization: 11.4%), and writing Internet comments intended to humiliate (perpetration: 11.5%; victimization: 24.4%). One reason for this lack of information is the absence of an instrument with good psychometric properties to measure CB adequately. Furthermore, the absence of an appropriate measure of CB in Mexico makes it difficult to conduct epidemiological studies on the causes, correlates, and consequences of this problem. Moreover, this absence precludes the comparisons between Mexico and other cultural contexts. Also, at the international level, the development of a valid and reliable measure is of vital importance for the development of research on CB (Slonje et al., 2013; Tokunaga, 2010). Although the instruments used in most of these studies have provided valuable insights into CB (for reviews, see Calvete et al., 2010; Tokunaga, 2010), these studies have several important limitations. First, as noted by Tokunaga (2010), “[CB] occurrence is most frequently operationalized in the form of one- or two-item measures based on dichotomous choice, yes/no responses, following a supplied definition of traditional bullying, cyberbullying, or both” (p. 283). Consequently, the range of CB behaviors analyzed by most of the scales has been very limited. Second, most of the available instruments have focused on assessing either victimization or perpetration of CB. Given that it has been found that perpetration and victimization tend to occur together (Estévez et al., 2010; Kowalski & Limber, 2007), it is important to have tools to measure them together. Finally, there are few measures of CB that are reported with their psychometric properties, and there are even fewer instruments that have been validated in different cultural contexts. This limitation has made it difficult to compare results from different cultures (Tokunaga, 2010).

Psychometric Properties of the Cyberbullying Questionnaire (CBQ)

67

The Cyberbullying Questionnaire (CBQ; Calvete et al., 2010; Estévez et al., 2010) is one of the first instruments that has been developed to measure the wide range of behaviors related to CB. This scale includes both perpetration of and victimization by CB and has been demonstrated to have adequate psychometric properties. A CBQ study on a sample of 1,431 Spanish adolescents has shown that this instrument has good construct validity and adequate reliability. Although these data are promising, much remains to be done concerning the psychometric properties of the CBQ. To date, only preliminary data about the convergent validity of this scale are available, that is, the extent to which the CBQ is related to other variables that would be expected to show a relationship. Furthermore, no study has examined the psychometric properties of the CBQ in other cultural contexts and whether the factor structure of the scale is invariant across different cultures.

FACTORS ASSOCIATED WITH Cyberbullying Gender has been one of the most frequently studied variables in research on bullying and CB. A review of the literature shows that gender plays an important role in traditional forms of bullying (Borg, 1999; Seals & Young, 2003). In addition, studies show that boys are more often both the aggressors and the victims of traditional bullying (Boulton & Underwood, 1992; Keiley, Bates, Dodge, & Pettit, 2000; Keltikangas-Järvinen, 2002). However, in the case of CB, the results are more complex. Although most studies have found that boys are more likely than girls to perpetrate CB (Calvete et al., 2010; DeHue, Bolman, & Völlink, 2008; Li, 2006), some studies have found either that there are no gender differences (Hinduja & Patchin, 2008; Willard, 2006) or even that the perpetration of CB is more common among girls (e.g., Kowalsky & Limber, 2007). Regarding victimization, most studies have found no differences between boys and girls (Hinduja & Patchin, 2008; Williams & Guerra, 2007; Wolak, Mitchell, & Finkelhor, 2007). However, some studies have also reported that more girls are victims of CB than boys (DeHue et al., 2008; Estévez et al., 2010; Smith et al., 2006). Thus, together, the results regarding gender have been mixed. In this regard, several studies suggest that a dependence on particular measures of aggressive behavior may confound the emergence of gender differences (Calvete & Cardeñoso, 2005). It has been suggested that girls use relational modes of aggression (e.g., spreading rumors), whereas boys use overt modes of aggression (e.g., physical aggression; Björkqvist, 1994; Owens, Shute, & Slee, 2000). The Internet and mobile phones may provide the perfect setting for the perpetration of some forms of relational aggression, and thus the results on the differences between boys and girls are not as conclusive as in the case of traditional bullying. Another factor that appears to contribute to the realization of CB is justification of the use of violence. Justification of the use of violence has been widely associated with the perpetration of various interpersonally aggressive behaviors (Espelage & Swearer, 2003; Huesmann & Guerra, 1997; Muñoz-Rivas, Gámez-Guadix, Fernández-González, & González, 2011). Regarding CB, Williams and Guerra (2007) found that the moral justification of bullying increased the likelihood of conducting CB behaviors. Similarly, Calvete et al. (2010) reported that the justification of violence, as understood as the appropriate use of aggression over a range of situations, significantly increased the likelihood of the perpetration of CB. Recently, Heirman and Walrave (2012) found that justification of CB behaviors was related to the behavioral intention to perpetrate them, which, in turn, was significantly associated with perpetration.

68

Gámez-Guadix et al.

A third variable associated with bullying and CB is impulsivity. Several studies suggest that impulsivity contributes to the exercise of bullying behaviors and CB (Fanti, Demetriou, & Hawa, 2012). Although impulsivity has been associated with both perpetration and victimization (e.g., Fanti & Kimonis, 2012), research suggests that impulsivity is associated more strongly with the former than the latter (Bosworth, Espelage, & Simon, 1999; Olweus, 1994). In this regard, O’Brennan, Bradshaw, and Sawyer (2009) found that among adolescents, the “bully/victims” and the “only bullies” showed similar levels of impulsivity to each and significantly higher levels of impulsivity than their peers who were not involved in bullying. However, the “only victims” had higher levels of impulsivity than their peers who were not involved in bullying but lower levels of impulsivity than the “only bullies” and the “bully/victims.” Finally, CB has been associated with negative effects on the psychological adjustment of the victims, especially with the emergence of depressive symptoms (Garaigordobil, 2011; Mitchell, Ybarra, & Finkelhor, 2007; Tokunaga, 2010). These psychological problems appear to be similar to those of traditional bullying (Mason, 2008), and the effects of CB may persist long term as is the case for traditional bullying (Chapell et al., 2006). Depressive symptoms appear to be greater in adolescents who have been either victims only or victims and perpetrators of CB than in adolescents who were either perpetrators or neither aggressors nor victims (Estévez et al., 2010).

THE PRESENT STUDY Given the paucity of validated instruments for measuring CB, the first objective of this study was to analyze the psychometric properties of the CBQ among Mexican adolescents. This objective included an analysis of the factor structure, the internal consistency, and the convergent validity of the CBQ. Because there are no previous data on the CBQ among Mexican adolescents, it is not only important to evaluate its psychometric properties but also to analyze whether the specified factor structure is valid in a different culture than the one in which the questionnaire was initially developed (Spain). Thus, we analyzed whether different factor structures for CBQ fit the Mexican data. Next, we analyzed whether the factor structure of the scale was equivalent across two different samples, the Mexican and the Spanish. Regarding the convergent validity, we studied the associations between the CBQ and other variables that we expected to be related to the CBQ on the basis of previous empirical evidence. Toward this aim, we analyzed the relationship between CB and the presence of justification of CB, impulsivity, and depressive symptoms. First, in consistency with previous studies (Calvete et al., 2010, Williams & Guerra, 2007), we expected to find that a greater justification of CB is associated with higher scores on the CBQ. Because the justification of violence has been more closely linked to perpetration than to victimization, we hypothesized a stronger relationship between justification and perpetration of CB than between justification and victimization by CB (Calvete et al., 2010; Williams & Guerra, 2007). Second, we expected to find a relationship between impulsivity and CB. Based on the research on traditional bullying, we expected to find a stronger relationship between impulsivity and perpetration than between impulsivity and CB victimization (e.g., O’Brennan et al., 2009). Finally, we expected to find a positive relationship between CB and the presence of depressive symptoms, with a stronger relationship in the case of victimization than in the case of perpetration (Mitchell et al., 2007; Patchin & Hinduja, 2006; Ybarra & Mitchell, 2004).

Psychometric Properties of the Cyberbullying Questionnaire (CBQ)

69

The final objective of this study was to analyze gender differences in the prevalence of CB. Because most studies have found that the perpetration of CB is more common in boys than in girls (e.g., DeHue et al., 2008), we hypothesized that the prevalence of perpetration would be higher in boys than in girls in our study. Although the results regarding victimization have been inconsistent, most studies have found no differences between boys and girls (for review, see Tokunaga, 2010). Therefore, we expected to find that the prevalence of victimization among Mexican adolescents would be similar for boys and girls.

METHOD Participants The study sample consisted of 1,491 Mexican adolescents (52.4% male and 47.6% female), with a mean age of 14.51 years (SD 5 1.57, range 5 12–18). The adolescents came from 45 classrooms that were randomly selected from three schools in Mexico City using a cluster-sampling method, stratified by grade level. From each grade level (three grades in secondary education and two grades in high school), a sample of nine classrooms was selected, and all students in each selected classroom were evaluated. The schools were selected in various areas of the city to reflect different socioeconomic backgrounds. The sampling error was 2.5%. The 65.6% of the students were enrolled in secondary school and 34.4% were enrolled in high school. Regarding the educational level of their parents, most parents had completed secondary education (mothers: 25.0%; fathers: 26.3%), high school or vocational training (mothers: 40.7%; fathers: 40.1%), or were graduates or engineers (mothers: 19.4%; fathers: 22.9%). The adolescents in the sample made extensive use of new technologies: 94.1% used the Internet to search for information, 91.3% used the Internet to be on Facebook, and 85.2% used the Internet to check e-mail. In addition, 84.0% talked on a cell phone and 85.0% sent text messages. Participants in the Spanish sample, only used for the multigroup analysis, were 1,008 adolescents (55.7% girls, 40.0% boys, and 4.3% did not indicate gender) with a mean age of 15.23 years (SD 5 1.4). Students came from 24 classrooms from 10 schools in the province of Bizkaia, Spain. The schools were chosen from all centers of Bizkaia using simple random sampling. The schools that gave permission for the study allowed us to evaluate various courses of secondary compulsory education and high school. Of the participants, 77.4% were in secondary compulsory education and 22.6% were in high school. This sample was collected between November and December 2011. The Mexican and the Spanish samples significantly differed in participants’ age, t(2,473) 5 12.54, p , .001, which was higher in the Mexican sample, and participants’ sex x2(1) 5 7.76, p , .01, with more boys among the Mexican participants.

Measures Cyberbullying Questionnaire (CBQ). The CBQ (Calvete et al., 2010) was composed of two different scales, one for measuring CB perpetration and one for measuring CB victimization. The initial version of the questionnaire consisted of 16 items on perpetration and 11 items on victimization. In this study, we reviewed the initial versions of the CBQ and eliminated two of the items in each scale to avoid redundancies in the content and to

70

Gámez-Guadix et al.

adapt the questionnaire to new ways of using new technologies. For example, we merged two items from the older version (“Sending threatening or insulting messages by e-mail” and “Sending threatening or insulting messages by cell phone”) into a more general single item in the revised version (i.e., “Sending threatening or insulting messages”) because threatening or insulting messages are usually sent over multiple media (social networks, e-mail, text messages, chats, smartphone applications, such as WhatsApp, etc.) and through any electronic device (e.g., smartphone, tablet, or computer). Thus, the revised CB perpetration subscale included 14 items. When tested with this subscale, the adolescent must indicate how often he or she had performed CB behaviors, such as sending threatening or insulting messages to other people or uploading or sending humiliating images, while on the Internet or on a cell phone. The victimization subscale included 9 items regarding the frequency with which adolescents have suffered different behaviors of CB. The participants were asked how often they had experienced different CB behaviors via the Internet or cell phone, such as receiving insulting or threatening messages or images of themselves that were humiliating while using the Internet or a cell phone. The response format used to assess how often each behavior had occurred as CB is as follows: 0 (never), 1 (1 or 2 times), 2 (3 or 4 times), or 3 (5 or more times). To adapt the scale to a Mexican linguistic context, we made several minor changes in the formulation of the items, replacing some words with alternatives appropriate to Mexico (e.g., “teléfono móvil” was replaced by “celular”). Next, the Spanish version of the questionnaire was pilot-tested to 126 Mexican adolescents who were enrolled in several courses to detect possible difficulties in the comprehension of the items. The content of each item was discussed with the adolescents to ensure that the items were ­understandable. Justification of Cyberbullying Scale. Five specific questions were included to assess whether adolescents justify the CB behaviors. The justification of CB is defined here as the set of ideas that maintains that the aggression is appropriate for various motives such as revenge (“Threatening or insulting another person who previously did those things to you, via cell phone or Internet”), restitution (“Recording or sending videos or pictures to post on the Internet of someone else who deserved it”, “Posting humiliating images or links on the Internet of a classmate who deserved it”), to laugh or make fun of someone (“Sending messages, e-mails, or pictures to someone else to mock or laugh at him or her”), or deliberately causing harm to someone (“Writing or sending jokes, rumors, gossip, or comments to ridicule a classmate”). The response format consisted of six categories ranging from 1 (never justified) to 6 (always justified). An exploratory factor analysis using the principal axis extraction method yielded a single factor (inspection of the Scree plot, eigenvalues greater than 1) that explained 63.99% of the variance, and all of the items had factor loadings greater than .56. The internal consistency of this scale was a 5 .85. Impulsivity. We used the Spanish adaptation of the Dysfunctional Impulsivity subscale of the Dickman Impulsivity Inventory (Chico, Tous, Lorenzo-Seva, & Vigil-Colet, 2003). This subscale consists of 12 items (e.g., “I often say and do things without considering the consequences”) with a response format of six categories ranging from 1 (never) to 6 (always). The Spanish version of this scale has been shown to have adequate construct validity and internal consistency in various studies (Chico et al., 2003). The reliability in the present study was a 5 .79. Depressive Symptoms. We used five items from the depression subscale (e.g., “Feeling no interest in things”) of the Spanish version of the Brief Symptom Inventory (BSI; Derogatis & Fitzpatrick, 2004; Pereda, Forns, & Peró, 2007). The psychometric properties

Psychometric Properties of the Cyberbullying Questionnaire (CBQ)

71

of the Spanish version of the BSI have been shown to have construct validity and good ­internal consistency (Pereda et al., 2007). The response format consisted of six categories ranging from 1 (not at all) to 6 (extremely). The internal consistency in this sample was a 5 .84.

Procedure The procedures for this research were explained to school administrators and to the Associations of Students’ Parents after informing them of the study’s objectives. After granting their permission, participants were informed of the anonymous and voluntary nature of the questionnaire, and they were given the opportunity to resolve any questions individually with the researcher in charge of the classroom. We asked the adolescents not to write their names anywhere on the questionnaires. The adolescents completed the questionnaires during their regular class schedules. The sessions lasted approximately 50 min. Only five adolescents (0.3%) refused to participate. The measurements were taken between May and June of 2011. The research was approved by the appropriate school board ethics committees.

RESULTS Psychometric Properties of the Cyberbullying Questionnaire Factor Structure. A series of confirmatory factor analyses were conducted using LISREL 8.8 (Jöreskog & y Sörbom, 2004). Because students were nested within classrooms, we computed standard errors and chi-square tests of model fit taking into account the nonindependence of data, using the Cluster option in LISREL. The structural equation models were tested using weighted least squares (WLS) estimation. Following the recommendations of several authors (Byrne, 2006; Hu & Bentler, 1999), the goodness of fit was evaluated with the non-normative fit index (NNFI), the comparative fit index (CFI), and the root mean square error of approximation (RMSEA). Generally, NNFI and CFI values of .90 or higher reflect a good fit. In addition, RMSEA values less than .05 indicate excellent fit, whereas values between .06 and .08 indicate acceptable fit. The hypothesized model consisted of two correlated factors, one for CB perpetration and one for CB victimization. The correlations between the measurement errors for the parallel items (i.e., pairs of items of perpetration and victimization sharing similar content) were allowed. For example, we estimated the correlation between the measurement error of Item 1 of perpetration (“Sending insulting or threatening messages to others”) and the measurement error of Item 1 of victimization (“Receiving threatening or insulting messages”), as undertaken in the measurement models for repeated measures. This way, we controlled the part of the variance that was expected to be common for these items (Little, Preacher, Selig, & Card, 2007). The solution obtained was satisfactory with good fit indices: x2 (220, N 5 1,491) 5 293, p , .001; NNFI 5 .98, CFI 5 .99, and RMSEA 5 .030 (95% CI: 0.027, 0.034). The model is presented in Figure 1. The loadings for each of the factors were statistically significant in all cases greater than .81 for perpetration and .59 for victimization. The correlation between perpetration and victimization was .71 (p , .001). We estimated three alternative factor models for comparison with the hypothesized model. Fit indices for the alternative models are presented in Table 1. The first alternative model was similar to the hypothesized model, but correlations between the measurement

72

Gámez-Guadix et al. Item 1

Item 1 Item 2

Item 2 Item 3 Item 4

.59

.81

.92

Item 3

.85

.98

Item 5 .79

.81

.93

Item 6

.71

.92 Item 7

.75

.98 Perpetration

.91 Item 8

.93

Victimization

Item 5

.99

.97 .94

Item 9

Item 4

.87

Item 6

.94 .91

Item 10

.86 Item 7

1 .96

Item 11

.95

Item 12 Item 8 Item 13 Item 14

Item 9

Figure 1.  Estimated model for the Cyberbullying Questionnaire (CBQ). Note. x2(220, N 5 1491) 5 293, p , .001; NNFI 5 .98, CFI 5 .99, RMSEA 5 0.030 [.027; .034].

errors for parallel items were constricted to zero. This imposition increased x2 significantly, Dx2(9, N 5 1,491) 5 101, p , .001, thereby indicating the appropriateness of allowing these correlations in the model. The second model consisted of two uncorrelated factors (perpetration and victimization). This model displayed poor fit indexes and increased x2 significantly, Dx2(1, N 5 1,491) 5 3,297, p , .001. The third model was onedimensional. This model also increased x2 significantly, Dx2(1, N 5 1,491) 5 17,300, p , .001. Thus, these comparisons between models indicate that victimization and perpetration are two different, but correlated, latent variables.

Equivalence of the Model Across Samples The next step was to evaluate the factor invariance of the CBQ across Mexican and Spanish students using multiple-group-covariance-structure analysis. First, the configural invariance of the model was tested to demonstrate that the pattern of fixed and free parameters

Psychometric Properties of the Cyberbullying Questionnaire (CBQ)

73

TABLE 1.  Goodness-of-Fit of the Hypothesized and Alternative Models x2

Df

NNFI

CFI

RMSEA

Hypothesized model. Two correlated factors, one for CB perpetration and one for CB victimization. Correlations between the measurement errors for parallel items were allowed.

293

220

.98

.99

.030 [.027–.034]

First alternative model. Two correlated factors, one for CB perpetration and one for CB victimization. Correlations between the measurement errors for parallel items were constricted to zero.

394

229

.98

.98

.034 [.031–.038]

3628

221

.80

.83

.11 [.100–.110]

17,380

221

.06

.18

.23 [.230–.240]

Second alternative model. Two uncorrelated factors (perpetration and victimization). Third alternative model. One-dimensional model.

was equivalent across subsamples, x2(440, N 5 2,499) 5 517, ns; NNFI 5 1, CFI 5 1, and RMSEA 5 .012 (95% CI 5 0.007; 0.016). Then, the invariance of the variances and covariances of the latent variables was tested. The change in x2 indicated that the overall pattern of variances and covariances was equivalent, Dx2(3, N 5 2,499) 5 3, ns. Finally, the invariance of the factor loadings was tested. The nonsignificant x2 ­increment indicated that the overall pattern of factor loadings is similar across Mexican and Spanish adolescents, Dx2(24, N 5 2,499) 5 25, ns. Internal Consistency. We calculated the internal consistency (Cronbach’s alpha) of the scales of perpetration and victimization. For perpetration, the internal consistency was a 5 .90, and for victimization, the internal consistency was a 5 .79. Convergent Validity. Convergent validity was assessed by analyzing the relationships of the CBQ scales with the scores of the other variables that we expected to be related. These results are presented in Table 2.

TABLE 2.  Correlations Among CB Perpetration and Victimization, Justification of CB, Impulsivity, and Depressive Symptoms Media

SD

1

1.  Cyberbullying perpetration

1.14

0.27

2.  Cyberbullying victimization

1.13

0.30

.62

3.  Justification of cyberbullying

1.30

0.59

.45

.31

4. Impulsivity

2.50

0.79

.26

.25

.15

5. Depression

1.74

0.78

.16

.27

.22

Note. All of the correlations are significant at the level **p , .01.

2

3

4

.29

74

Gámez-Guadix et al.

The justification of CB was associated with both perpetration and victimization, although the relationship was significantly higher with the former than with the latter (r 5 .45 for perpetration and victimization for r 5 .31, t 5 3.29, p ,. 01). In addition, the analysis also showed significant correlations between impulsivity and the CBQ scales. In this case, there are no significant differences between the correlations of impulsivityperpetration and impulsivity-victimization (r 5 .26 and r 5 .25, respectively). Lastly, the scale of depressive symptoms showed a significant relationship with both perpetration and victimization, although with a higher correlation for victimization (r 5 .27) than for perpetration (r 5 .16; t 5 22.07, p , .01).

Gender Differences in Cyberbullying The total percentage of students who reported that they had performed at least one CB behavior was 41.9%. This percentage was significantly higher for boys than for girls (males: 46.5%; females: 38.5%; x2[1, N 5 1,420] 5 7.67, p , .01). Regarding ­victimization, 44.5% of the adolescents reported having been victims of CB. In this case, we found no significant differences between males and females (males: 46.5%; females: 44.5%; x2[1, N 5 1,420] 5 0.54, ns). Table 3 presents the percentage of adolescents who performed each type of CB behavior. For perpetration, the percentages ranged between 2.0% and 18.2% for the girls and between 4.7% and 22.9% for the boys. Most of the analyzed behaviors were significantly more common among boys than girls (see Table 3). Regarding victimization (see Table 4), the prevalence of different types of victimization ranged between 2.2% and 19.3% for the girls and between 6.4% and 22.5% for the boys. In four out of the nine types of victimization, the levels of prevalence were significantly higher for boys than for girls; the other behaviors did not differ by gender. For both victimization and perpetration, the most common types of CB were “writing jokes, rumors, gossip, or ridiculous comments” and “sending threats or insults.”

DISCUSSION This research aimed to increase our knowledge about the measurement and the prevalence of CB. The study contributes to the Mexican adaptation of a reliable and valid instrument for CB, to expand the evidence on the variables associated with CB, and to examine the extent of this problem among adolescents. The results of this study provided data on the adequate factorial validity of the CBQ in a Mexican sample that were in line with previously reported data for a Spanish sample (Calvete et al., 2010; Estévez et al., 2010). Specifically, the findings of this study support a structure for the CBQ that consists of two distinct factors: perpetration and victimization. The results did not support either the unidimensional model or the two uncorrelated factors for the CBQ. However, even though taking into account the need to differentiate between these two factors, it is important to note that perpetration and victimization showed a high relationship with each other, suggesting the need to measure them together. This finding is consistent with results of previous studies that have found that perpetration and victimization often occur simultaneously (Estévez et al., 2010; Kowalski & Limber, 2007). Multigroup factor analyses provided additional support for the factor structure of the CBQ because the hypothesized model was equivalent for both the Mexican and the

Psychometric Properties of the Cyberbullying Questionnaire (CBQ)

75

TABLE 3.  Prevalence and Gender Differences for Cyberbullying Perpetration Females

Males

x2

10.7%

15.1%

6.01*

  2. Posting or sending humiliating images of classmates

6.6%

13.2%

17.06***

  3. Posting links of humiliating images to other people for them to see

7.3%

13.0%

12.50***

  4. Writing embarrassing jokes, rumors, gossip, or comments about a classmate on the Internet

18.2%

22.9%

4.93*

  5. Posting or sending links with rumors, gossip, etc. about a classmate to other people so they can read them

10.4%

12.8%

1.92

  6. Hacking to send messages by e-mail or social networks that could make trouble for the other person

6.0%

13.2%

20.70***

  7. Recording a video or taking pictures by cell phone while a group laughs and forces another person to do something humiliating or ridiculous

3.7%

10.6%

25.25***

  8. Posting or sending these images to be seen by other people

8.8%

13.4%

7.70**

  9. Recording a video or taking pictures by cell phone while someone hits or hurts another person

4.7%

10.3%

15.97***

10. Posting or sending these images to be seen by other people

5.1%

11.5%

18.81***

11. Broadcasting online other people’s secrets, compromising information or images

7.5%

11.3%

5.97*

10.1%

11.5%

0.75

13. Recording a video or taking cell phone pictures of classmates performing some type of behavior of a sexual nature

2.0%

4.7%

7.75**

14. Hanging or sending these images to be seen by other people

1.2%

5.8%

22.26***

  1.  Sending threatening or insulting messages

12. Deliberately excluding someone from an online group

Note. *p  .05, **p  .01, ***p  .001. Spanish sample. In addition, as is the case in the works that originally validated the CBQ, the perpetration and victimization scales showed adequate internal consistencies. Overall, the results provided additional support for the convergent validity of the CBQ. As expected, we found a significant positive relationship between CB and the justification of this type of aggressive behavior. Although the relationships of justification with perpetration and victimization were both significant, the relationship was stronger with perpetration (Heirman & Walrave, 2012). Moreover, as in the relationships between other violent behaviors and the attitudes that justify them (e.g., Muñoz-Rivas et al., 2011), the effect size of the relationship was moderate.

76

Gámez-Guadix et al.

TABLE 4.  Prevalence and Gender Differences for Cyberbullying Victimization Females

Males

x2

17.0%

17.9%

0.20

5.9%

11.0%

3. Writing embarrassing jokes, rumors, gossip, or comments about me on the Internet

19.3%

22.5%

2.18

4. Hacking me to send messages by e-mail or social networks that could be troublesome for me

14.2%

13.7%

0.03

5. Recording a video or taking pictures by cell phone while a group laughs and forces me to do something humiliating or ridiculous

3.1%

7.1%

13.47***

6. Recording a video or taking pictures by cell phone while someone hits or hurts me

2.2%

6.4%

14.84***

7. Broadcasting online secrets, compromising information or images about me

10.2%

9.6%

0.15

8.  Deliberately excluding me from an online group

11.3%

10.4%

0.25

2.2%

6.1%

1.  Receiving threatening or insulting messages 2. Posting on the Internet or sending humiliating images of me

9. Recording a video or taking cell phone pictures of me performing some type of behavior of a sexual nature

11.98***

13.38***

Note. ***p  .001. The relationship between impulsivity and CB was also significant, as expected. However, contrary to our hypothesis, we did not find differences in the size of relationship between impulsivity-perpetration and impulsivity-victimization (O’Brennan et al., 2009). However, the significant positive relationship between impulsivity and victimization is not an unexpected finding. In this regard, it has been noted that the lack of emotional and behavioral control that are characteristic of high levels of impulsivity may lead teens to engage in behaviors that capture the attention of the bullies and increase the likelihood of their own victimization (Fanti & Kimonis, 2012). Furthermore, the relationship between impulsivity and victimization is consistent with findings that peer victimization is more common in youth with attention deficit/hyperactivity disorder (ADHD), in which impulsivity is a defining characteristic (Bagwell, Molina, Pelham, & Hoza, 2001). Finally, as hypothesized, we found a relationship between depressive symptoms and the CBQ subscales. The relationship between victimization and depression was stronger than the relationship between perpetration and depression, although in both cases the relationship was significant (Estévez et al., 2010). These data are consistent with the body of prior research that indicates that CB is associated with poorer psychosocial adjustment, including depressive symptoms (e.g., Mitchell et al., 2007). The results also indicated a high prevalence of CB in Mexico. Approximately 42% of the adolescents said they had committed some type of CB, whereas 44% reported being victims of CB. Comparing these data with those obtained in international studies, we found that the level of prevalence was slightly higher in this study. For example, the literature review by Tokunaga (2010) indicated that most authors have reported levels of prevalence of CB

Psychometric Properties of the Cyberbullying Questionnaire (CBQ)

77

between 20% and 40%. However, when comparing the levels of p­ revalence for the individual behaviors of CB in the present work with those obtained in another study conducted in Mexico (Lucio, 2009), we found that the results in both studies were similar. For example, the most common behaviors in the study were receiving threats (perpetration: 22%; victimization: 19.6%) and spreading rumors affecting one’s reputation (perpetration: 10.7%; victimization: 15.4%) with levels of prevalence comparable to those obtained in this study. There are several possible explanations for the higher prevalence obtained in this study in comparison with the results in other countries. First, it is possible that the prevalence of CB in Mexico is higher than in other countries because of cultural and social factors (e.g., greater exposure of adolescents to violence in the community). Second, it is possible that because CB is a growing phenomenon, the high percentages reflect that growing problem. Third, the use of an instrument such as the CBQ, which includes a wide range of CB behaviors, could have made teens more likely to answer “yes” to any of the CB behaviors, increasing the level of prevalence as a consequence (Calvete et al., 2010). Regarding gender differences, more boys perpetrated CB behaviors than girls. These findings are consistent with previous studies on traditional bullying and CB (Keiley et al., 2000; Li, 2007; Slonje & Smith, 2007). In addition, 12 out of the 14 CB behaviors studied were higher for boys than for girls. Only two types of relational aggression did not differ between males and females (“Posting or sending links with rumors, gossip, etc. about a classmate to other people so they can read them” and “Deliberately excluding someone from an online group”). The greater tendency for boys to show externalizing behavior problems, including aggression (Negriff & Susman, 2011), might explain why more boys perpetrated CB than girls. Regarding victimization, the overall prevalence of victims of CB did not differ by gender (46.5% and 44.5% for boys and girls, respectively). This finding is consistent with most of previous studies, which found no differences between males and females in CB victimization (DeHue et al., 2008; Kowalski & Limber, 2007; Smith et al., 2006). However, it should be noted that in four out the nine CB behaviors studied, mainly those related to happy slapping (e.g., “record me on video or take pictures with the phone while someone hits or hurts me”), the prevalence was higher for males. Thus, the data indicate that boys and girls are affected similarly by CB but that boys seem to be more involved in CB behaviors involving recording compromising or humiliating situations and disseminating those images. This modality of CB involves being a victim of aggression in a real setting in addition to the electronic context. The greater tendency of boys to be bullied through physical aggression, as compared with girls (Tokunaga, 2010), might explain these findings.

Limitations and Future Research This study has several limitations that should be noted. First, the results are based on the self-reports of adolescents via questionnaires, a design feature that might have introduced some bias into the information (e.g., increasing the shared method variance). Future studies should include the reports of others (e.g., peers, teachers, parents) and should use other assessment techniques, such as interviews. Second, this study provides evidence of some of the psychometric properties of the instrument (i.e., factorial validity, convergent validity, and reliability), but future studies should provide additional data on other types of validity (e.g., predictive) and the test–retest reliability of the scale. Finally, although the sample is large, it is not representative of the Mexican population, so one must be cautious in generalizing the findings. Future studies should replicate these findings with additional samples.

78

Gámez-Guadix et al.

CONCLUSIONS In summary, the findings of this study indicate that CB behaviors are highly prevalent and seem to be a common feature of interpersonal interactions during adolescence. Furthermore, the results suggest that the CBQ is a useful instrument that has adequate psychometric properties among adolescents. The CBQ has several advantages: it measures both perpetration and victimization; it assesses a wide range of CB behaviors; and it is brief and easy to complete. The CBQ is particularly suited to settings in which it can be employed to evaluate CB behaviors before and after the application of prevention programs. In terms of research, the CBQ can be used to analyze temporal relationships between the potential predictors and consequences of CB and CB behaviors using a longitudinal design, which is one of the main challenges in future research concerning this problem (Tokunaga, 2010).

REFERENCES Bagwell, C. L., Molina, B. S. G., Pelham, W. E., Jr., & Hoza, B. (2001). Attention-deficit hyperactivity disorder and problems in peer relations: Predictions from childhood to adolescence. Journal of the American Academy of Child & Adolescent Psychiatry, 40, 1285–1292. Björkqvist, K. (1994). Sex differences in physical, verbal, and indirect aggression: A review of recent research. Sex Roles, 30, 177–188. Borg, M. G. (1999). The extent and nature of bullying among primary and secondary schoolchildren. Educational Research, 41, 137–153. 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. Boulton, M. J., & Underwood, K. (1992). Bully/victim problems among middle school children. British Journal of Educational Psychology, 62, 73–87. Byrne, B. M. (2006). Structural equation modeling with EQS. Basic concepts, applications, and programming (2nd ed.). Hills Dale, NJ: Lawrence Erlbaum Associates. Calvete, E., & Cardeñoso, O. (2005). Gender differences in cognitive vulnerability to depression and behavior problems in adolescents. Journal of Abnormal Child Psychology, 33(2), 179–192. Calvete, E., Orue, I., Estévez, A., Villardón, L., & Padilla, P. (2010). Cyberbullying in adolescents: Modalities and aggressors’ profile. Computers in Human Behavior, 26, 1128–1135. Chapell, M. S., Hasselman, S. L., Kitchin, T., Lomon, S. N., MacIver, K. W., & Sarullo, P. L. (2006). Bullying in elementary school, high school, and college. Adolescence, 41, 633–648. Chico, E., Tous, J. M., Lorenzo-Seva, U., & Vigil-Colet, A. (2003). Spanish adaptation of Dickman’s impulsivity inventory: Its relationship to Eysenck’s Personality Questionnaire. Personality and Individual Differences, 35, 1883–1892. DeHue, F., Bolman, C., & Völlink, T. (2008). Cyberbullying: Youngsters’ experiences and parental perception. CyberPsychology & Behavior, 11, 217–223. Derogatis, L. R., & Fitzpatrick, M. (2004). The SCL-90-R, the Brief Symptom Inventory (BSI), and the BSI-18. Hillsdale, NJ: Lawrence Erlbaum Associates. Espelage, D. L., & Swearer, S. M. (2003). Research on school bullying and victimization: What have we learned and where do we go from here? School Psychology Review, 32, 365–383. Estévez, A., Villardón, L., Calvete, E., Padilla, P., & Orue, I. (2010). Adolescentes víctimas de cyberbullying: Prevalencia y características. Psicología Conductual, 18, 73–89. Fanti, K. A., Demetriou, A. G., & Hawa, V. V. (2012). A longitudinal study of cyberbullying: Examining risk and protective factors. European Journal of Developmental Psychology, 9, 168–181.

Psychometric Properties of the Cyberbullying Questionnaire (CBQ)

79

Fanti, K. A., & Kimonis, E. R. (2012). Bullying and victimization: The role of conduct problems and psychopathic traits. Journal of Research on Adolescence, 22, 617–631. Garaigordobil, M. (2011). Prevalencia y consecuencias del cyberbullying: Una revisión. International Journal of Psychology and Psychological Therapy, 11, 233–254. Heirman, W., & Walrave, M. (2012). Predicting adolescent perpetration in cyberbullying: An application of the theory of planned behavior. Psicothema, 24, 614–620. Hinduja, S., & Patchin, J. W. (2008). Cyberbullying: An exploratory analysis of factors related to offending and victimization. Deviant Behavior, 29, 129–156. Hu, L., & Bentler, P. M. (1999). Cut off criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55. 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. Jöreskog, K. G., & y Sörbon, D. (2004). LISREL 8.8 for Windows [Computer Software]. Lincolnwood, IL: Scientific Software International. Keiley, M. K., Bates, J. E., Dodge, K. A., & Pettit, G. S. (2000). A cross-domain growth analysis: Externalizing and internalizing behaviors during 8 years of childhood. Journal of Abnormal Child Psychology, 28, 161–179. Keltikangas-Järvinen, L. (2002). Aggressive problem-solving strategies, aggressive behavior, and social acceptance in early and late adolescence. Journal of Youth and Adolescence, 31, 279–287. Kiriakidis, S. P., & Kavoura, A. (2010). Cyberbullying: A review of the literature on harassment through the internet and other electronic means. Family & Community Health, 33, 82–93. Kowalsky, R. M., & Limber, S. P. (2007). Electronic bullying among middle school students. Journal of Adolescent Health, 41, S22–S30. Li, Q. (2006). Cyberbullying in schools: A research of gender differences. School Psychology International, 27, 157–170. Li, Q. (2007). New bottle but old wine: A research of cyberbullying in schools. Computers in Human Behavior, 23, 1777–1791. Li, Q. (2008). A cross-cultural comparison of adolescents’ experience related to cyberbullying. Educational Research, 50, 223–234. Little, T. D., Preacher, K. J., Selig, J. P., & Card, N. A. (2007). New developments in latent variable panel analyses of longitudinal data. International Journal of Behavioral Development, 31, 357–365. Lucio, L. A. (2009). Agresores escolares en el ciberespacio; el cyberbullying en preparatorias Mexicanas. Retrieved from http://www.convivenciaescolar.net/wp/wp-content /uploads/2009/10/ARTICULO_ECUADOR.pdf Mason, K. L. (2008). Cyberbullying: A preliminary assessment for school personnel. Psychology in the Schools, 45, 323–348. Mitchell, K. J., Ybarra, M., & Finkelhor, D. (2007). The relative importance of online victimization in understanding depression, delinquency, and substance use. Child Maltreatment, 12, 314–324. Muñoz-Rivas, M. J., Gámez-Guadix, M., Fernández-González, L., & González, M. P. (2011). Validation of the Attitudes about Aggression in Dating Situations (AADS) and the Justification of Verbal/Coercive Tactics Scale (JVCT) in Spanish adolescents. Journal of Family Violence, 26, 575–584. Negriff, S., & Susman, E. J. (2011). Pubertal timing, depression, and externalizing problems: A framework, review, and examination of gender differences. Journal of Research on Adolescence, 21, 717–746. O’Brennan, L. M., Bradshaw, C. P., & Sawyer, A. L. (2009). Examining developmental differences in the social-emotional problems among frequent bullies, victims, and bully/victims. Psychology in the Schools, 46, 100–115. Olweus, D. (1994). Bullying at school: Basic facts and effects of a school based intervention ­program. Journal of Child Psychology and Psychiatry, 35, 1171–1190.

80

Gámez-Guadix et al.

Owens, L., Shute, R., & Slee, P. (2000). “Guess what I just heard!”: Indirect aggression among teenage girls in Australia. Aggressive Behavior, 26, 67–83. Patchin, J. W., & Hinduja, S. (2006). Bullies move beyond the schoolyard: A preliminary look at cyberbullying. Youth Violence and Juvenile Justice, 4, 148–169. Pereda, N., Forns, M., & Peró, M. (2007). Dimensional structure of the Brief Symptom Inventory with Spanish college students. Psicothema, 19, 634–639. Seals, D., & Young, J. (2003). Bullying and victimization: Prevalence and relationship to gender, grade level, ethnicity, self-esteem, and depression. Adolescence, 38, 735–748. Slonje, R., & Smith, P. K. (2007). Cyberbullying: Another main type of bullying? Scandinavian Journal of Psychology, 49, 147–154. Slonje, R., Smith, P. K., & Frisén, A. (2013). The nature of cyberbullying, and strategies for prevention. Computers in Human Behavior, 29, 26–32. Smith, P., Mahdavi, J., Carvalho, M., & Tippett, N. (2006). An investigation into cyberbullying, its forms, awareness and impact, and the relationship between age and gender in cyberbullying (Research Brief No. RBX03-06). London, United Kingdom: DfES. Tokunaga, R. S. (2010). Following you home from school: A critical review and synthesis of research on cyberbullying victimization. Computers in Human Behavior, 26, 277–287. Willard, N. (2006). Cyberbullying and cyberthreats. Eugene, OR: Center for Safe and Responsible Internet Use. Williams, K. R., & Guerra, N. G. (2007). Prevalence and predictors of internet bullying. Journal of Adolescent Health, 41, 14–21. Wolak, J., Mitchell, K., & Finkelhor, D. (2007). Unwanted and wanted exposure to online pornography in a national sample of youth Internet users. Pediatrics, 119, 247–257. World Internet Project in Mexico. (2013). Estudio 2012 de hábitos y percepciones de los mexicanos sobre Internet y diversas tecnologías asociadas. Retrieved from http://www.wip.mx/ Ybarra, M. L., & Mitchell, K. J. (2004). Youth engaging in online harassment: Associations with caregiver–child relationships, Internet use, and personal characteristics. Journal of Adolescence, 27, 319–336. Acknowledgment. This research was partially supported by Ministerio de Economía y Competitividad (Spanish Government) grant PSI2012-31550. Correspondence regarding this article should be directed to Manuel Gámez-Guadix, PhD, University of Deusto, Bilbao, Vizcaya Spain. E-mail: [email protected]

Violence and Victims, Volume 28, Number 5, 2013

Do Networking Activities Outside of the Classroom Protect Students Against Being Bullied? A Field Study With Students in Secondary School Settings in Germany Gerhard Blickle, PhD University of Bonn

James A. Meurs, PhD University of Mississippi

Christine Schoepe, Dipl-Psych University of Cologne Research has shown that having close relationships with fellow classmates can provide a buffer for students against bullying and the negative outcomes associated with it. But, research has not explicitly examined the potential benefits of social networking behaviors outside of the classroom for those who could be bullied. This study addresses this gap and finds that, although a bullying climate in the classroom increases overall bullying, students high on external networking activities did not experience an increase in the bullying they received when in a classroom with a high bullying climate. However, the same group of students reported the largest degree of received bulling under conditions of a low bullying climate. We discuss the implications of our results and provide directions for future research.

Keywords: networking; bullying; secondary school; classroom climate

I

n recent years, many schools have moved from developing prevention programs aimed at reducing aggression to programs specifically focused on preventing bullying (see for reviews Merrell, Gueldner, Ross, & Isava, 2008; Ttofi, Farrington, & Baldry, 2008). Most antibullying programs have demonstrated weak or nonsignificant effects (Bauer, Lozano, & Rivara, 2007; Jenson & Dieterich, 2007; Smith, Schneider, Smith, & Ananiadou, 2004). One review (i.e., Ttofi et al., 2008) found several characteristics of effective programs, some of which suggest the importance of school and classroom climate. Other researchers also have indicated that climate is important to bullying in schools (Peeters, Cillessen, & Scholte, 2010; Unnever & Cornell, 2003). In addition, some research has shown that social networks within the classroom can provide positive outcomes for students, particularly for those facing bullying from other students (e.g., Davies, 1984; Pellegrini, Bartini, & Brooks, 1999). But, not all social

© 2013 Springer Publishing Company

81

82

Blickle et al.

networks provide protection from bullying (Champion, Vernberg, & Shipman, 2003; Espelage, Holt, & Henkel, 2003; Holt & Espelage, 2007; Pellegrini et al., 1999). Research also indicate that bullying victims tend to be socially isolated and lacking social skills (Craig, 1998; Smokowski & Kopasz, 2005), and one study found that (low) peer preference, not peer rejection, promotes bullying when in a bullying climate (i.e., Sentse, Scholte, Salmivalli, & Voeten, 2007). We argue that it is not the social network that is of help to the student, but networking behaviors that allow the student to reduce the likelihood of being bullied. In addition, we were unable to locate research that specifically examined the networking behaviors that students direct to those outside of their class of fellow students. Thus, our research question asks whether these external networking behaviors of students help to prevent them from being bullied when in a high bullying climate classroom.

THEORETICAL BACKGROUND School climate has been conceptualized as the interaction quality and frequency among adults and students (Bandyopadhyay, Cornell, & Konold, 2009). In school-related research, the influence of climate has lead researchers (e.g., O’Moore, 2000) to suggest that a climate that ignores or reinforces bullying may contribute to children’s reluctance to tell parents or school authorities when they have been victimized. A positive school climate has been shown to be related to decreased bullying and victimization (Guerra, Williams, & Sadek, 2011). Other research has shown that students attending schools that are less supportive (e.g., Gendron, Williams, & Guerra, 2011; Nansel et al., 2001), have a more negative climate (e.g., Kasen, Berenson, Cohen, & Johnson, 2004; Olweus, 1993), or have victimization (e.g., Hong & Eamon, 2012) are more likely to engage in bullying behavior. Moreover, bullying can be characterized as the aggression of a group composed of the bully and his or her supporters against the victim and his or her supporters (Salmivalli, Huttunen, & Lagerspetz, 1997). Thus, the bullying climate among that group of individuals (e.g., a class) would seem to be an influential component of behavior in such a situation. Also, students can take on various roles regarding the bullying behavior, such as a defender of the victim (Sainio, Veenstra, Huitsin, & Salmivalli, 2011) or a bystander (see Salmivalli, 2010). These roles also suggest the importance of the climate among these peer witnesses. Unnever and Cornell (2003) characterized the bullying culture of a school as the extent to which there are widespread perceptions that bullying behavior will not be interrupted or face intervention. However, overall, as suggested by Peeters et al. (2010), despite its potential importance, the influence of the classroom context in peer relations research has been underappreciated and provides avenues for future research. We argue that, much like a school’s culture and climate influence bullying and victimization, each school’s classroom would have an impact on the behavior of the students in that class. We believe this is especially true because the climate of the group of one’s closest peers is more proximal than the climate of the overall school. Of the studies examining students’ social connections and bullying, few have differentiated between established social networks and networking activities. For example, Huitsing, Veenstra, Sainio, and Salmivalli (2012) examined bullying and victimization networks by conducting a social network analysis. Others have demonstrated that the social position and acceptance are important factors in bullying or joining bullying groups (e.g., Olthof & Goossens, 2008; Witvliet et al., 2010). Although these lines of research are

Networking and Bullying

83

valuable and they reinforce the importance of the social aspects of bullying, one’s position in a social network is not the same as networking activities. Social networking activities are behaviors focused on the construction and maintenance of informal relationships (Ferris et al., 2007; Wolff, Moser, & Grau, 2008). These behaviors are not the structures of (already established) social networks, as represented by constructs such as social capital (Coleman, 1988; Putnam, 1993). In the context of a school, a student could have a large social network including students, teachers, and administrators. These social relationships might not have been established through networking activities, but instead through, for example, the social connections a teacher has with a student’s parents (e.g., attending the same church or synagogue) or siblings (e.g., a sports teammate of a sister). Certainly, networking activities lead to the creation of social networks and social capital (Thompson, 2005). But, social networks are not completely due to social ­networking activities, especially for children. School-aged children have a much more limited social environment than adults. In addition, victims of bullying tend to have social adjustment difficulties (Nansel et al., 2001; Smokowski & Kopasz, 2005) and, over time, victimization becomes a social role (Salmivalli, 2010). Also, because bullies are often perceived as powerful, student witnesses may be more likely to avoid assisting the low-status victims (Juvonen & Galvan, 2008). This could explain why victims have been found to be socially isolated (Craig, 1998) and have more difficulties in their friendships than nonvictims (Parker & Asher, 1993). Furthermore, these findings might indicate why victims report decreased peer support (Seeds, Harkness, & Quilty, 2010) in spite of the fact that they also report valuing social support more than others (Demaray & Malecki, 2003). It has been suggested that although victims in some cases report high levels of social support, they might feel unable to use it (Holt & Espelage, 2007). We argue that these findings in the literature are partly because of the ability to successfully engage in the behavior that creates and maintains social networks, allowing youth one way to prevent the occurrence of bullying. In other words, the poor relationships and social isolation of victimized students could be caused by social networking deficiencies, and those who possess such social adroitness are able to prevent bullying. This possibility has rarely been examined empirically. For example, one study found that in classrooms with a high bullying climate, bullying was more strongly related to (low) peer preference than peer rejection (Sentse et al., 2007). This could suggest that scholars should examine networking behaviors rather than network presence or perceived peer support, because the social influence of networking behaviors could build peer preference, whereas one’s network position might only prevent peer rejection. Furthermore, we were unable to find any prior examination of the networking behaviors of students outside of the classroom in relation to bullying. However, the results of some studies could offer insight into the potential effects. Unexpectedly, Boulton (1999) did not find a relationship between peer popularity (i.e., group size of companions) and victimization. The author suggested that one possibility was that the characteristics of one’s friends matters more to victimization than the number of friends/companions. In support of this suggestion, one study (i.e., Hodges, Malone, & Perry, 1997) found that the link between behavioral risk and victimization was stronger for those with physically weaker friends than for those with physically stronger friends. Similarly, we believe it is possible that external networking activities allow students to escape from their classroom’s bullying climate without receiving as much bullying from fellow classmates. In addition, it should make a difference whether a student is socially passive or socially active, proactively

84

Blickle et al.

seeking new social contacts (Thompson, 2005). Those who proactively seek new social contacts outside their class should have easier access to higher levels of social support because they can mobilize more support from others (Lin, 2008). Holt and Espelage (2007) found that high levels of social support were related to increased anxiety and depression for victims of bullying, suggesting that these children possibly felt unable to use their support effectively. Moreover, they argued that future research needs to assess social support strategies more comprehensively. In their study, Holt and Espelage’s measure of social support included and combined support from mothers and close friends. The measure also did not isolate support within the classroom from the support that was external to the classroom, and it did not separate socially passive and socially proactive children. We believe that Holt and Espelage’s finding of high support leading to poorer outcomes for victims also could be because of this lack of partitioning of the data. Finally, Holt and Espelage also called for research to consider how school climate factors relate to peer relations, and scholars have begun to address this call (e.g., Goldstein, Young, & Boyd, 2008; Hong & Espelage, 2012; see Salmivalli, 2010). We contend that our study addresses both of these future research directions desired by Holt and Espelage (2007). The proactive creation and maintenance of external social networks (i.e., social networking activities outside of the classroom) could be considered a strategy for reducing the likelihood of being bullied when in a classroom with a high bullying climate. Furthermore, the Holt and Espelage measure of social support included support both in the classroom and outside of the classroom. Thus, we believe our results could provide at least a partial explanation for the findings of their research, because our study specifically examines students networking behaviors external to the classroom environment. Clearly, previous studies have shown that having close relationships with peers inoculates students from being bullied, and we believe this will extend to those situations where bullying is most frequent (i.e., high bullying climate in the class). In other words, a student who is proactively networking with those outside of the classroom can engage in these activities to prevent the consequences (i.e., received bullying) of a high bullying climate within the classroom. In line with prior research, we argue that, when in a classroom climate high on bullying, students who engage in a high degree of networking outside of the classroom will be able to use it as a way of escaping from the bullying behavior that is happening in their class. It also is likely to build peer liking and preference among fellow students (Sentse et al., 2007). Whereas, for those students who engage in little networking outside of their classroom peers, a greater bullying climate in their classroom will lead to increased experiences of being bullied. Thus, we offer the following hypothesis: Students’ networking activities outside of the classroom will moderate the classroom bullying climate—degree of being bullied relationship, such that those high on networking will not experience an increase in bullying under conditions of a heightened bullying climate. However, for students low on networking, a greater bullying climate will lead to increased received bullying.

STUDY CONTEXT This study was conducted in Germany. Wolke, Woods, Stanford, and Schulz (2001) found that 16% of students were victims and 9% were bullies in a German sample. In another study from Germany, Kasper and Heinzelmann-Arnold (2008) reported that approximately one out of six students between 10 and 19 years, with boys and girls demonstrating similar

Networking and Bullying

85

rates, was a victim of bullying behavior. In this study, the following definition of bullying guided the research (Olweus, 1996). A student is being bullied, when another student or several other students •  say mean and hurtful things or make fun of him or her or call him or her mean and hurtful names •  completely ignore or exclude him or her from their group of friends or leave him or her out of things on purpose •  hit, kick, push, shove around, or lock him or her inside a room •  tell lies or spread false rumors about him or her or send mean notes and try to make other students dislike him or her •  and other hurtful things like that. When we talk about bullying, these things happen repeatedly and it is difficult for the student being bullied to defend himself or herself. We also call it bullying, when a student is teased repeatedly in a mean and hurtful way. But we don’t call it bullying when the teasing is done in a friendly and playful way. Also, it is not bullying when two students of about equal strength or power argue or fight. (Olweus, 1996, p. 1)

Our research objective is to test whether networking activities can be a protective factor against being bullied. Networking activities necessitate advanced social skills, which are more frequently found among the age group of 10th- to 13th-grade students. Thus, because student networking activities play a key (i.e., moderating) role in our research design, the data collection took place in four secondary schools.

METHOD Participants and the Context of Our Study The data collection took place in four secondary schools, three state schools, and one private school in Germany. Three of the schools were located in a metropolitan region and one was in a rural region. Data was collected from 10th- to 13th-grade students, and the data collection process occurred in the last week before summer holidays. For students younger than the age of 16 years, parents were asked in advance for permission to let their children take part in the study. No such student was denied participation in the study by her or his parents. The survey questionnaires were handed out in class by the research team in the presence of a teacher. Of the 738 students who were in class, all worked on the questionnaire and returned it to the research team in the same hour. There were 706 students who returned complete questionnaires. Thus, the return rate of complete questionnaires was 95.6%. Of the students participating, 310 (42.1%) were males and 424 (57.5%) females; three students did not report their gender. The mean age of the students was 16.97 (SD 5 1.04) years. In German schools, there are regular breaks between classes, allowing students to socialize with students from other classes in the school yard. Furthermore, not only do many students live in the same neighborhood and ride busses to school together (providing more opportunity for social interaction), there also are many occasions where students interact with students in other classes in curricular/extracurricular groups (e.g., music/band practice, theatrical performances, chess, choir, and computer clubs). In short, the German school systems provides ample opportunities for students to socially engage and network with students outside of their classroom, and we believe this presents a valuable context for examining the effects of this external social networking.

86

Blickle et al.

Measures Received Bullying. The degree to which a student has been bullied in the course of the previous month was assessed with the victim subscale of the Bullying Participant Roles Survey (BPRS; Summers, Demaray, & Becker, 2010) by the students. The BPRS is based on 12 items originally created by Olweus (1996) and Salmivalli and Voeten (2004). The BPRS covers the following dimensions of classroom bullying: verbal bullying (insulting, offending; example item: “I have been called mean names”), relational bullying (leave someone out on purpose, ignoring someone; example item: “I have been purposely left out of something”), physical bullying (punching, slapping, pinching, poking; example item: “I have been pushed around, punched, or slapped”), and object-related bullying (taking things away from someone; example item: “I have had things taken away from me”). The items were translated into German, and the back translation was checked by two native American speakers with expertise in educational settings. To also cover cyber-­ bullying, two items were added (i.e., “Someone in my class insulted me, dissed me or threatened me on email, text messaging or telephone calls.” “Someone in my class insulted me, dissed me, or threatened me using pictures or videoclips, on websites or in chatrooms.”). The complete scale consisted of 14 items and used a 5-point response format (1 5 never, 2 5 1–2 times, 3 5 3–4 times, 4 5 5–6 times, 5 5 7 times or more). Cronbach’s alpha in the present research was a 5 .89. Bullying Climate in Class. To assess the bullying climate in class, a scale consisting of 33 items was used. Thirty-one of the items were taken from the Smob (Schueler Mobbing) questionnaire (Kasper & Heinzelmann-Arnold, 2008). The Smob is an adaptation of Leymann Inventory of Psychological Terror (LIPT, Leymann, 1990) to be used with students in school settings. The Smob items were transferred from the individual level to the level of classes (e.g., “I have been punched” to “How often did it happen in your class in the last three weeks that another student has been punched?”). Furthermore, two items were added to cover the dimension cyber-bullying (i.e., “How often did it happen in your class in the last three weeks that another student received insulting or dissing emails, text messages or telephone calls?” “How often did it happen in your class in the last three weeks that another student was insulted, dissed or threatened using pictures or videoclips, on websites or in chatrooms?”). The questionnaire used a 4-point Likert-type scale (no/never, once, more than once, more than one time per week). Smob was reported to have demonstrated good psychometric qualities (Kasper, 2000). Cronbach’s alpha in the present research was a 5 .94. Networking Outside of the Classroom. The workplace networking scale by Wolff and Moser (2006; Wolff, Schneider-Rahm, & Forret, 2011) was adapted to the school setting from the perspective of an individual student. Three studies conducted by Wolff and Moser had provided evidence for high psychometric qualities (Wolff & Moser, 2009). The Wolff and Moser networking scale comprises six facets, namely, building versus maintaining versus using contacts inside and outside the institution. The combination of these facets leads to six subscales of networking: Building internal contacts (e.g., “When there is somebody new in school, I introduce myself personally”), maintaining internal contacts (e.g., “I discuss problems in school with students from other classes.”), using internal contacts (e.g., “I ask students from other classes to gather information for me.”), building external contacts (e.g., “I’m building contacts out of school to have contact persons there as well.”), maintaining external contacts (e.g., “I use contacts out of school to get some advice.”), and using external contacts (e.g., “I exchange hints and advices with friends

Networking and Bullying

87

from other schools.”), where internal refers to internal to the school and external refers to external to the school. The Networking Scale comprises 44 items and uses a 4-point Likert-type scale (ranging from never/very seldom to very often/always). Cronbach’s alpha in the present research was a 5 .92. Negative Affectivity State in Class. Our measures of bullying climate in class and of received bullying neither use the term “bully” nor did they define or describe bullying. Consequently, to distinguish bullying from kidding and joking, we assessed student’s negative mood while in the classroom, with the understanding that bullying climate and received bullying should (positively) correlate with a negative mood if he/ she is interpreted by the students as actual bullying behavior rather than mere joking or having fun. In recent research, two broad, general factors (i.e., positive and negative affect) have emerged as the dominant dimensions of emotional experience. To measure these factors, Watson, Clark, and Tellegen (1988) developed the Positive and Negative Affect Schedule (PANAS), which consists of 10-item scales for PA and NA, respectively. Each item is rated on a 5-point Likert-type scale (ranging from 1 5 very slightly or not at all to 5 5 extremely). To use the PANAS with German students, the scale needed to be translated. Krohne, Egloff, Kohlmann, and Tausch (1996) created a German version of the PANAS. Just like the English version (Watson et al., 1988), the German adaptation of the PANAS provides high psychometric qualities (Krohne et al., 1996). In this study, we only used the Negative Affect scale, and our Cronbach’s alpha was a 5 .94. Control Variables. Research has demonstrated the impact of students’ gender (Schäfer, Korn, Werner, & Crick, 2006), age (Perry, Kusel, & Perry, 1988), and neuroticism (Byrne, 1994; Coyne, Chong, Seigne, & Randall, 2003; Mynard & Joseph, 1997; Slee & Rigby, 1993) on students’ victim status in bullying situations. Therefore, we controlled for these variables in our analyses. Male was coded 0, female was coded 1, and age was measured in years. To assess neuroticism, the BFI-K (Rammstedt & John, 2005) self-rating Neuroticism scale was used. It comprises four items on a 5-point Likert-type sale ranging from 1 5 strongly disagree to 5 5 strongly agree. Two previous studies demonstrated high psychometric qualities of the BFI-K (Rammstedt & John, 2005). Cronbach’s alpha in the present research was a 5 .74. Because networking is highly correlated with extraversion (Ferris et al., 2008; Wolff & Moser, 2006), we additionally controlled for extraversion. If the hypothesized effects of students’ networking activities emerge above and beyond extraversion, we can exclude extraversion as rival explanation of the effects of networking activities. To assess extraversion, the BFI-K (Rammstedt & John, 2005) self-rating Extraversion scale was used. It comprises four items on a 5-point Likert-type scale ranging from 15 strongly disagree to 5 5 strongly agree. As mentioned earlier, the BFI-K provides high psychometric qualities (Rammstedt & John, 2005). To control for sequential arrangements of the scales in the survey, the whole survey used five different permutations. Finally, we controlled for the four different schools in which the data were collected.

Data Analysis First, for each student we aggregated the ratings of the bullying climate in class of his classmates and dropped the respective student’s climate ratings. This provided an objective measure of bullying class climate for each student independent of the respective student.

88

Blickle et al.

Second, a hierarchical moderated multiple regression analysis (Cohen, Cohen, West, & Aiken, 2003) was conducted to examine the moderating role of networking activities on the effects of bullying climate in class on the degree of being bullied in class. The networking scale and the bullying in class climate scale were centered prior to creating the interactive term. In the first step, we included the control variables, namely, age, gender, neuroticism, and extraversion. In the second step, we entered the main effects of student’s networking activities and bullying climate in class. In the third step, the interactive term of networking activities 3 bullying climate in class was entered. If the interaction effect of networking activities 3 bullying climate in class is significant and the interaction plot demonstrates a positive effect of high networking activities in the presence of increased bullying climate, it would confirm our research hypothesis.

RESULTS The intraclass correlation indices (ICC) of bullying class climate were (ICC[1] 5 .12) and (ICC[1, k] 5 .66). Therefore, based on LeBreton and Senter (2008), for each student we aggregated the ratings of the bullying climate in class of his/her classmates and dropped the respective student’s climate ratings. This provided an objective measure of bullying class climate for each student independent of the respective student. The means, standard deviations, correlations, and coefficient alpha (a) internal consistency reliability estimates of the study variables are reported in Table 1. As discussed earlier, to test that our bullying received and bullying climate measures were perceived as bullying rather than mere kidding or joking, we assessed student negative affective mood while in the classroom. Thus, the correlations between received bullying, bullying climate in class, and negative affectivity were calculated. Negative affectivity in class correlated positively with bullying received (r 5 .34, p , .01) and with bullying climate in class (r 5 .10, p , .05). These findings clearly demonstrate that the content of the bullying items were negatively evaluated and were not perceived merely as kidding. A hierarchical moderated multiple regression analysis (Cohen et al., 2003) was conducted to test our hypothesis. The hypothesis stated that networking activities would moderate the relationship between bullying climate in class and received bullying. The data analysis confirmed this expectation. The interaction term of networking activities 3 bullying climate had a significant beta-weight (b 5 2.54, p , .01) and explained 1% additional variance in the criterion (see Table 2). In a covariance analysis (Winer, 1970) that combines nominal variables (e.g., the different schools) and interval variables (e.g., networking activities, bullying climate) as predictors, we reran the above analyses with the four different schools as additional control variable in the first step. However, the different schools had no impact on the results of these analyses. The interaction term of the centered networking activities and bullying climate variables had the same significant beta-weight. Therefore, we excluded the nominal school variable from further analyses. The form of the “networking 3 bullying climate in class” interaction was illustrated according to the procedure proposed by Cohen et al. (2003). Three levels of networking were plotted: at one standard deviation below the mean, at the mean, and at one standard deviation above the mean.

Networking and Bullying

TABLE 1.  Means, Standard Deviations, Coefficient Alpha Reliabilities, and Correlations of the Variables Variables

M

SD

1

1. Gender

0.58

0.494

—

16.97

1.039

2.08*

3. Neuroticism

2.82

0.918

.30**

2.02

(.74)

4. Extraversion

3.57

0.914

.12**

2.03

2.24**

(.80)

5. Negative affectivity state in class

1.68

0.606

2.17**

.07*

.21**

2.05

(.84)

6. Bullying climate in class

1.77

0.219

2.14**

2.27**

2.05

.10*

7. Networking activity

2.25

0.406

.22**

.04

8. Received bullying

1.35

0.516

2.24**

2.04

2. Age

2

3

4

5

6

7

8

—

2.05 .04 .08*

.34** 2.05

2.03 .37**

(.94) 2.17** .13**

(.92) 2.05

(.89)

Note. N 5 706; Gender: 0 5 male, 1 5 female. *p , .05. **p , .01.

89

90

Blickle et al.

TABLE 2.  Hierarchical Regression Analysis on Received Bullying Criterion Variable 5 Degree of Being Bullied Block

Predictors

1

Gender

2.321**

Age

2.028

2

3

B

Neuroticism

.103**

Extraversion

.018

Bullying climate in class

.217*

Networking activity

.013

Bullying climate 3 Networking

R2

DR2

.090

.090

.098

.008

.108

.010

2.539**

Note. N 5 706; Gender: 0 5 male, 1 5 female. *p , .05. **p , .01. Figure 1 presents the significant “networking 3 class climate” interaction effect. As expected, for students low in networking (i.e., one SD below mean) higher levels of bullying climate in class were associated with higher levels of students’ victim status. The slope of this regression line was positive and significant (b 5 .41, p , .001). For students high on networking (i.e., one SD above mean), bullying climate in class was not associated with students’ victim status. Thus, the results generally provided support for the research hypothesis. However, under conditions of a low bullying climate, the students high in networking activities reported the highest amount of received bullying.

DISCUSSION Overall, our results supported our hypothesis. Unlike their classmates with low external networking, students with greater networking external to their class did not experience increased bullying when in a high bullying climate. But, those low on networking received the greatest amount of bullying when the bullying climate in class was high. Our study also indicates the importance of classroom climate, because, on the whole, received bullying increased for students when in a high bullying climate. It may be that the climate of a student’s class is more relevant to student outcomes than a school’s climate, which may include the climate of other students, teachers, and staff with whom the students do not regularly interact. This result supports the suggestion that the characteristics of the student’s social network might be important when considering student outcomes. The ability of students to connect with those outside of their class could be particularly helpful when their

Networking and Bullying

91

1.82

Networking High

1.80

Medium * Low **

1.78

Degree of Being Bullied

1.76 1.74 1.72 1.70 1.68 1.66 1.64 1.62 Low

Medium Bullying Climate in Class

High

Figure 1.  Interaction of bullying climate in class and students’ external networking activity on degree of being bullied (received bullying), controlling for gender, age, neuroticism, and extraversion. N 5 706. *p , .05. **p , .01.

classroom’s climate is pervasive in bullying behaviors, such that it prevents them from receiving as much bullying as their classmates. The precise explanations for the use of external networks has yet to be uncovered, and educational scholars should investigate the underlying mechanisms of what makes a student’s social network more or less helpful when in the presence of a destructive (e.g., bullying) classroom environment. Overall, our results indicate that researchers should consider narrow constructs when assessing student behavioral outcomes, such as external social networks and classroom (not overall school) climate. Our finding that in low bullying climates, those high on networking outside of the class received the most bullying was unexpected. Boulton (1999) showed that boys (but not girls) who engaged in social sedentary activity (e.g., conversation; “hanging around”; and chatting) had an increased likelihood of being subsequent victims. However, another study (i.e., Schwartz, Dodge, & Coie, 1993) found that male victims of bullying initiated fewer conversations before they became victims. Neither of these studies took into consideration the climate of the school or classroom, and they used a sample of children from ages 6–9 years. The effects of social networking for older children might be very different than for younger children, and the students in our study had an average age of nearly 17 years.

92

Blickle et al.

Most children who are victimized by bullies also have been found to be rejected by their peers (Perry et al., 1988; Salmivalli, 2010). It could be that those high on external networking are external networkers because of rejection by classmates. Furthermore, as researchers have suggested, if only a small percentage of fellow students defend the victims of bullies, over time, victims might be inclined to seek support from outside of the classroom. A study by Cowie (2000) found that victims in the age range of 11–14 years tended to use a peer support system and most of those who accessed it found it to be helpful. This finding could suggest that, longitudinally, those who are bullied may seek support from those outside of their peers. However, longitudinal research found that peer rejection was related to antisocial and bullying behavior about 2 years later (Dishion, Patterson, Stoolmiller, & Skinner, 1991). Another study demonstrated that over a 3-year period, social isolation was related to future bullying and victimization (Rubin & Mills, 1988). Clearly, over time, the relationships between peer rejection, networking outside of the classroom, and bullying and victim status become complex. Future longitudinal research is needed to better explain how the relationships of these constructs develop and change over time.

Strengths, Limitations, and Directions for Future Research Strengths include the collection of data from both students and their peers, minimizing concerns about the presence of common method bias (Podsakoff, Mackenzie, Lee, & Podsakoff, 2003). Furthermore, the sample was fairly large thereby minimizing b-errors (Cohen, 1992). Finally, all scales used to operationalize study variables of interest demonstrated strong psychometric properties. Our research has some limitations. Namely, our measure of bullying behavior used self-reports, and it could be argued that this measures the perceptions of being bullied by others. Reports from other students or teachers of a person being bullied also would be relevant and could be examined by future research. One perceived limitation of our study could be the low means and variances of bullying climate in the classroom and of received bullying, indicating that little bullying occurred in our sample. However, because of the extreme nature of the verbal and physical violence associated with bullying (e.g., punching, insulting, or threatening someone), it is a rare occurrence, when compared to other types of interactions. Furthermore, from a methodological standpoint, if our sample was range restricted in bullying behavior, it would decrease the probability of discovering our interaction effect, suggesting that our significant interaction results are a conservative estimate of the true effects. Also, because we did not measure networking inside the classroom, we cannot rule out the possibility that students who are high external networkers also are high networkers in the classroom. However, our finding that, in low bullying climates, high external networking students received the greatest amount of bullying would seem to suggest there is not a strong correlation between internal and external networking. In addition, it has been suggested that whole-school prevention programs like the Olweus Bullying Prevention Program (Olweus et al., 2007) might be the best at reducing the shame in backing down from a fight, establishing norms of nonviolence, and teaching students prosocial/nonaggressive ways of handling interpersonal conflict (Ttofi et al., 2008). Our results indicating the influence of classroom bullying climate on bullying behavior supports these suggestions. Moreover, the findings suggest that bullying prevention programs should consider prevention at the classroom level. Future research should

Networking and Bullying

93

seek to establish the effectiveness of such whole-school or climate-related bullying prevention programs at both school and classroom levels of analysis.

Practical Implications Most antibullying programs at school have demonstrated, at best, weak effects (Bauer et al., 2007). One prevention program that seems to be quite effective (i.e., Olweus Bullying Prevention Program; Olweus et al., 2007) is focused on the whole school, and its effectiveness could be because it focuses on changing the climate of the school. This study also suggests that practitioners also should consider the climate of the classes when addressing the issue of bullying at school. In their systematic review, Ttofi et al. (2008) found several elements of effective antibullying programs, and some of these also indicate the importance of classroom climate (e.g., classroom rules and classroom management). Moreover, the review demonstrated these programs to be more effective with older children and to yield better results when there was increased intensity and duration of the program. Practitioners should consider the results of this and other reviews of antibullying programs when forming school policies and procedures (e.g., directing programs to older, rather than younger children). As noted by Espelage, Bosworth, and Simon (2000) concerning bullying behavior, school counselors should think of bullying climate and bullying received on a continuum, rather than trying to classify, for instance, if a student is a victim or a bully. Appreciating the range of bullying climates and behaviors also supports the development of more robust and comprehensive bullying prevention programs. At the individual level, prior studies suggest that students have improved outcomes when they have close relationships with fellow classmates (Davies, 1984; Pellegrini et al., 1999). The results of our research indicate that parents, teachers, and administrators should encourage students to also maintain relationships with those outside of their classroom when bullying is a common occurrence in their classroom. In classrooms high on bullying, these social networking behaviors by students likely help them to build greater acceptance from their peers (Sentse et al., 2007). These social influence actions should also assist students when they are in environments low on bullying, as they progress through the educational system. Moreover, the effectiveness of social skills has been demonstrated to help adults adapt to the workplace (Ferris et al., 2007). Thus, the development of social networking skills can provide long-term benefits to students as they move into adulthood.

CONCLUSION The goal of the present research was to improve our understanding of how student’s networking activities outside of the classroom moderate the relationship between the climate of bullying in the classroom and the bullying received by that student. We found in a crosssectional design that students high on external networking did not experience an increase in the bullying they received when in a classroom with a high bullying climate. However, students reported the largest degree of received bulling under conditions of a low bullying climate. Future research should use predictive designs, thereby providing constructive replication and extension of the present findings.

94

Blickle et al.

REFERENCES Bandyopadhyay, S., Cornell, D. G., & Konold, T. R. (2009). Validity of three school climate scales to assess bullying, aggressive attitudes, and help seeking. School Psychology Review, 38, 338–355. Bauer, N. W., Lozano, P., & Rivara, F. P. (2007). The effectiveness of the Olweus bullying prevention program in public middle schools: A controlled trial. Journal of Adolescent Health, 40, 266–274. Boulton, M. J. (1999). Concurrent and longitudinal relations between children’s playground behavior and social preference, victimization, and bullying. Child Development, 70, 944–954. Byrne, B. J. (1994). Bullies and victims in school setting with reference to some Dublin schools. Irish Journal of Psychology, 15, 574–586. Champion, K., Vernberg, E., & Shipman, K. (2003). Nonbullying victims of bullies: Aggression, social skills, and friendship characteristics. Journal of Applied Developmental Psychology, 24, 535–551. Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155–159. Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression correlation analysis for the behavioral sciences. Mahwah, NJ: LawrenceErlbaum Associates. Coleman, J. S. (1988). Social capital in the creation of human capital. American Sociological Review, 94(Suppl.), S95–S120. Cowie, H. (2000). Bystanding or standing by: Gender issues in coping with bullying in English schools. Aggressive Behavior, 26, 85–97. Coyne, I., Chong, P. S.-L., Seigne, E., & Randall, P. (2003). Self and peer nominations of bullying: An analysis of incident rates, individual differences and perceptions of the working environment. European Journal of Work and Organizational Psychology, 12, 209–228. Craig, W. M. (1998). The relationship among bullying, victimization, depression, anxiety, and aggression in elementary school children. Personality and Individual Differences, 24, 123–130. Davies, B. (1984). Life in the classroom and playground. London, United Kingdom: Routledge & Kegan Paul. Demaray, M. K., & Malecki, C. K. (2003). Perceptions of the frequency and importance of social support by students classified as victims, bullies, and bully/victims in an urban middle school. School Psychology Review, 32, 471–489. Dishion, T. J., Patterson, G. R., Stoolmiller, M., & Skinner, M. L. (1991). Family, school, and behavioral antecedents to early adolescent involvement with antisocial peers. Developmental Psychology, 27, 172–180. Espelage, D. L., Bosworth, K., & Simon, T. R. (2000). Examining the social context of bullying behaviors in early adolescence. Journal of Counseling and Development, 78, 326–333. Espelage, D. L., Holt, M. K., & Henkel, R. R. (2003). Examination of peer group contextual effects on aggressive behavior during early adolescence. Child Development, 74, 205–220. Ferris, G. R., Blickle, G., Schneider, P. B., Kramer, J., Zettler, I., Solga, J., . . . Meurs, J. A. (2008). Political skill construct and criterion-related validation: A two-study investigation. Journal of Managerial Psychology, 23, 744–771. Ferris, G. R., Treadway, D. C., Perrewé, P.L., Brouer, R. L., Douglas, C., & Lux, S. (2007). Political skill in organizations. Journal of Management, 33, 290–320. Gendron, B. P., Williams, K. R., & Guerra, N. G. (2011). An analysis of bullying among students within schools: Estimating the effects of individual normative beliefs, self-esteem, and school climate. Journal of School Violence, 10, 150–164. Goldstein, S. E., Young, A., & Boyd, C. (2008). Relational aggression at school: Associations with school safety and social climate. Journal of Youth and Adolescence, 37, 641–654. Guerra, N. G., Williams, K. R., & Sadek, S. (2011). Understanding bullying and victimization during childhood and adolescence: A mixed methods study. Child Development, 82, 295–310. Hodges, E. V., Malone, M. J., & Perry, D. G. (1997). Individual risk and social risk as interacting determinants of victimization in the peer group. Developmental Psychology, 33, 1032–1039.

Networking and Bullying

95

Holt, M. K., & Espelage, D. L. (2007). Perceived social support among bullies, victims, and bullyvictims. Journal of Youth and Adolescence, 36, 984–994. Hong, J. S., & Eamon, M. K. (2012). Students’ perceptions of unsafe schools: An ecological systems analysis. Journal of Child and Family Studies, 21, 428–438. Hong, J. S., & Espelage, D. L. (2012). A review of research on bullying and peer victimization in school: An ecological system analysis. Aggression and Violent Behavior, 17, 311–322. Huitsing, G., Veenstra, R., Sainio, M., & Salmivalli, C. (2012). “It must be me” or “It could be them?”: The impact of the social network position of bullies and victims on victims’ adjustment. Social Networks, 34(4), 379–386. http://dx.doi.org/10.1016/j.socnet.2010.07.002 Jenson, J. M., & Dieterich, W. A. (2007). Effects of a skills-based prevention program on bullying and bully victimization among elementary school children. Prevention Science, 8, 285–296. Juvonen, J., & Galvan, A. (2008). Peer influence in involuntary social groups: Lessons from research on bullying. In M. Prinstein & K. Dodge (Eds.), Peer influence processes among youth (pp. 225–244). New York, NY: Guilford Press. Kasen, S., Berenson, K., Cohen, P., & Johnson, J. G. (2004). The effects of school climate on changes in aggressive and other behaviors related to bullying. In D. L. Espelage & S. Swearer (Eds.), Bullying in American schools (pp. 187–210). Mahwah, NJ: Lawrence Erlbaum Associates. Kasper, H. (2000). Schülermobbing—tun wir was dagegen! Der Smob—Fragebogen mit Anleitung und Auswertungshilfe und Materialien für die Schulentwicklung [Bullying students—let’s do something against it. The Smob-manual]. Lichtenau-Scherzheim, Germany: AOL Verlag. Kasper, H., & Heinzelmann-Arnold, I. (2008). Schülermobbing—tun wir was dagegen! [Students being bullied—What do we do to prevent it?]. Buxtehude, Germany: AOL-Verlag. Krohne, H. W., Egloff, B., Kohlmann, C.-W., & Tausch, A. (1996). Untersuchungen mit einer deutschen Version der “Positive and Negative Affect Schedule” [Investigations with the German version of the “Positive and Negative Affect Schedule” (PANAS)]. Diagnostica, 42, 139–156. LeBreton, J. M., & Senter, J. L. (2008). Answers to 20 questions about interrater reliability and interrater agreement. Organizational Research Methods, 11, 815–852. Leymann, H. (1990). Handbog för användning av LIPT—formuläret för kartläggning av risker för psykiskt väld i arbetsmiljön [The LIPT questionnaire—a manual]. Stockholm, Sweden: Violen. Lin, N. (2008). A network theory of social capital. In D. Castiglione, J. W. Van Deth, & G. Wolleb (Eds.), The handbook of social capital (pp. 50–69). Oxford, United Kingdom: Oxford University Press. Merrell, K. W., Gueldner, B. A., Ross, S. W., & Isava, D. M. (2008). How effective are school bullying intervention programs? A meta-analysis of intervention research. School Psychology Quarterly, 23, 26–42. Mynard, H., & Joseph, S. (1997). Bully/victim problems and their association with Eysenck’s personality dimensions in 8- to 13-year-olds. British Journal of Educational Psychology, 67, 51–54. Nansel, T. R., Overpeck, M., Pilla, R. S., Ruan, W. J., Simons-Morton, B. G., & Scheidt, P. (2001). Bullying behaviors among US youth: Prevalence and association with psychosocial adjustment. Journal of the American Medical Association, 285, 2094–2100. Olthof, T., & Goossens, F. (2008). Bullying and the need to belong: Early adolescents’ bullyingrelated behavior and the acceptance they desire and receive from particular classmates. Social Development, 17, 24–46. Olweus, D. (1993). Bullying at school: What we know and what we do. Malden, MA: Blackwell. Olweus, D. (1996). The revised Olweus Bully/Victim Questionnaire for Students. Bergen, Norway: University of Bergen. Olweus, D., Limber, S. P., Flerx, V. C., Mullin, N., Riese, J., & Snyder, M. (2007). Olweus Bullying Prevention Program: Schoolwide guide. Center City, MN: Hazelden. O’Moore, M. (2000). Critical issues for teacher training to counter bullying and victimization in Ireland. Aggressive Behavior, 26, 99–111.

96

Blickle et al.

Parker, J. G., & Asher, S. R. (1993). Friendship and friendship quality in middle childhood: Links with peer group acceptance and feelings of loneliness and social dissatisfaction. Developmental Psychology, 29, 611–621. Peeters, M., Cillessen, A. H., & Scholte, R. H. (2010). Clueless or powerful? Identifying subtypes of bullies in adolescence. Journal of Youth and Adolescence, 39, 1041–1052. Pellegrini, A. D., Bartini, M., & Brooks, F. (1999). School bullies, victims, and aggressive victims: Factors relating to group affiliation and victimization in early adolescence. Journal of Educational Psychology, 91, 216–224. Perry, D. G., Kusel, S. J., & Perry, I. C. (1988). Victims of peer aggression. Developmental Psychology, 24, 807–814. Podsakoff, P. M., Mackenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88, 879–903. Putnam, R. (1993). The prosperous community: Social capital and public life. American Prospect, 4, 35–42. Rammstedt, B., & John, O. P. (2005). Kurzversion des Big Five Inventory (BFI-K): Entwicklung und Validierung eines ökonomischen Inventars zur Erfassung der fünf Faktoren der Persönlichkeit. Diagnostica, 51, 195–206. Rubin, K. H., & Mills, R. S. L. (1988). The many faces of social isolation in childhood. Journal of Consulting and Clinical Psychology, 56, 916–924. Sainio, M., Veenstra, R., Huitsin, G., & Salmivalli, C. (2011). Victims and their defenders: A dyadic approach. International Journal of Behavioral Development, 35, 144–151. Salmivalli, C. (2010). Bullying and the peer group: A review. Aggression and Violent Behavior, 15, 112–120. Salmivalli, C., Huttunen, A., & Lagerspetz, K. M. (1997). Peer networks and bullying in schools. Scandinavian Journal of Psychology, 38, 305–312. Salmivalli, C., & Voeten, M. (2004). Connections between attitudes, group norms, and behaviour in bullying situations. International Journal of Behavioral Development, 28, 246–258. Schäfer, M., Korn, S., Werner, N., & Crick, N. (2006). “. . . und es wird immer schlimmer!” Zur Unterscheidung von Bullying und Viktimisierung in der weiterführenden Schule [“. . . and it is always getting worse!” Researching on bullying and victimization in secondary school settings]. (Forschungsbericht Nr. 183). München, Germany: Ludwig-Maximilians-Universität, Department Psychologie, Institut für Pädagogische Psychologie. Schwartz, D., Dodge, K. A., & Coie, J. D. (1993). The emergence of chronic peer victimization in boys’ play groups. Child Development, 64, 1755–1772. Seeds, P. M., Harkness, K. L., & Quilty, L. C. (2010). Parental maltreatment, bullying, and adolescent depression: Evidence for the mediating role of perceived social support. Journal of Clinical & Adolescent Psychology, 39, 681–692. Sentse, M., Scholte, R., Salmivalli, C., & Voeten, M. (2007). Person-group dissimilarity in involvement in bullying and its relation with social status. Journal of Abnormal Child Psychology, 35, 1009–1019. Slee, P. T., & Rigby, K. (1993). The relationship of Eysenck’s personality factors and self-esteem to bully/victim behavior in Australian schoolboys. Personality and Individual Differences, 14, 371–373. Smith, J. D., Schneider, B. H., Smith, P. K., & Ananiadou, K. (2004). The effectiveness of wholeschool antibullying programs: A synthesis of evaluation research. School Psychology Review, 33, 547–560. Smokowski, P. R., & Kopasz, K. H. (2005). Bullying in school: An overview of types, effects, family characteristics, and intervention strategies. Children & Schools, 27, 101–110. Summers, K. H., Demaray, M. K., & Becker, L. D. (2010). Assessment of bystander participant roles in bullying situations. Paper presented at the Annual Convention of the National Association of School Psychologists, Denver, CO.

Networking and Bullying

97

Thompson, J. A. (2005). Proactive personality and job performance: A social capital perspective. Journal of Applied Psychology, 90, 1011–1017. Ttofi, M. M., Farrington, D. P., & Baldry, A. C. (2008). Effectiveness of programmes to reduce bullying. Stockholm, Sweden: Swedish National Council for Crime Prevention. Unnever, J. D., & Cornell, D. G. (2003). The culture of bullying in middle school. Journal of School Violence, 2, 5–27. Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54(6), 1063–1070. Witvliet, M., Olthof, T., Hoeksma, J. B., Goossens, F. A., Smits, M., & Koot, H. (2010). Peer group affiliation of children: The role of perceived popularity, likeability, and behavioral similarity in bullying. Social Development, 19, 285–303. Winer, B. J. (1970). Statistical principles in experimental design. London, United Kingdom: McGraw-Hill. Wolff, H.-G., & Moser, K. (2006). Entwicklung und Validierung einer Networkingskala. Diagnostica, 52, 161–180. Wolff, H.-G., & Moser, K. (2009). Effects of networking on career success: A longitudinal study. Journal of Applied Psychology, 94, 196–206. Wolff, H.-G., Moser, K., & Grau, A. (2008). Networking: Theoretical foundations and construct validity. In J. Deller (Ed.), Readings in applied organizational behavior from the Lüneburg Symposium (pp. 101–118). Mehring, Germany: Rainer Hampp. Wolff, H.-G., Schneider-Rahm, C., & Forret, M. L. (2011). Adaptation of a German multidimensional networking scale into English. European Journal of Psychological Assessment, 27, 244–250. Wolke, D., Woods, S., Stanford, K., & Schulz, H. (2001). Bullying and victimization of primary school children in England and Germany: Prevalence and school factors. British Journal of Psychology, 92, 673–696. Correspondence regarding this article should be directed to Gerhard Blickle, Arbeits-, Organisationsund Wirtschaftspsychologie, Institut fuer Psychologie, Universitaet Bonn, Kaiser-Karl-Ring 9, 53111 Bonn, Germany. E-mail: [email protected]

Violence and Victims, Volume 29, Number 6, 2014

The Differential Impacts of Episodic, Chronic, and Cumulative Physical Bullying and Cyberbullying: The Effects of Victimization on the School Experiences, Social Support, and Mental Health of Rural Adolescents Paul R. Smokowski, PhD, MSW Caroline B. R. Evans, MSW Katie L. Cotter, MSW University of North Carolina at Chapel Hill North Carolina Academic Center for Excellence in Youth Violence Prevention Few studies have examined the impacts of past, current, and chronic physical bullying and cyberbullying on youth, especially in rural settings. This study augments this scant literature by exploring the school experiences, social support, and mental health outcomes for rural, middle school youth. The participants for this 2-year longitudinal study were 3,127 youth from 28 middle schools. Participants were classified as nonvictims, past victims (i.e., victimized during Year 1 but not Year 2), current victims (i.e., victimized during Year 2 but not Year 1), and chronic victims (i.e., victimized during both Year 1 and Year 2). Findings illustrated that chronic victimization resulted in the lowest levels of school satisfaction, social support, future optimism, and self-esteem. Chronic victims also reported the highest levels of school hassles, perceived discrimination, peer rejection, anxiety, depression, and externalizing behaviors. In terms of episodic victimization, current year victimization was associated with worse outcomes than past year victimization. Implications and limitations were discussed.

Keywords: bullying victimization; middle school; rural; adolescents

B

ullying is a pervasive problem in the United States, typically categorized into five forms: physical (e.g., hitting, kicking), verbal (e.g., teasing, name-calling), social (e.g., excluding, rumor spreading), extortion (e.g., asking for money), and cyber (e.g., sending harmful electronic messages; Tsang, Hui, & Law, 2012). Olweus’s (1993) seminal definition of bullying focused on power imbalance, intent, and repetition. Subsequent researchers added a fourth dimension in focusing on provocation (Frisen, Holmqvist, & Oscarsson, 2008).

© 2014 Springer Publishing Company99

100

Smokowski et al.

A national survey of 6th through 10th graders indicated that about 30% of students reported involvement in bullying as a bully or victim in the current semester (Nansel et al., 2001). In the Centers for Disease Control and Prevention’s Youth Risk Behavior Surveillance System (Centers for Disease Control and Prevention [CDC], 2013), 20% of high school youth in 2009 and 2011 reported being bullied at school in the past year. This rate was higher for females than for males, decreased from 9th to 12th grade, and was highest for White and mixed-race adolescents. Another national survey found that 28% of adolescents reported bullying victimization with variation by subtype (National Center for Educational Statistics, 2011). Rates for involvement in physical and verbal bullying at least once in the past 2 months were 20.8% and 53.6%, respectively (Wang, Iannotti, & Nansel, 2009). Rates of cyberbullying vary from 5% to 40%, depending on the age group and definition of cyberbullying (Hinduja & Patchin, 2007). The prevalence of bullying in rural areas may be even higher. In a study of 192 rural 3rd through 8th grade students, 82% reported being bullied at least once over the past 3 months (Dulmus, Theriot, Sowers, & Blackburn, 2004). The behavioral, emotional, and physical consequences of bullying are well documented (Nansel, Craig, Overpeck, Saluja, & Ruan, 2004; Nansel et al., 2001; Smokowski & Holland, 2005). In one study, victims and bully/victims had the highest number of adjustment problems, whereas bullies had the lowest (Gini, 2008). Victims reported feeling powerless, excluded, and unsafe (Smokowski & Holland, 2005). Bullying victimization has been associated with decreased self-esteem (Guerra, Williams, & Sadek, 2011), low social competence (Nation, Vieno, Perkins, & Santinello, 2008), poor social and emotional adjustment (Nansel et al., 2004; Nansel et al., 2001), and low school attendance (Gastic, 2008). Being victimized also increased internalizing problems (Sweeting, Young, West, & Der, 2006), nervousness, (Gini, 2008), peer relationship problems, loneliness (Nansel et al., 2004), and social withdrawal (Cho, Hendrickson, & Mock, 2009). The effects of cyberbullying victimization are similar to traditional bullying and include feelings of anger, sadness, powerlessness, fear, and low self-esteem (Hoff & Mitchell, 2009). Victims of cyberbullying are at an increased risk of using alcohol and drugs, skipping school, receiving poor grades, experiencing in-person bullying, and suffering from health problems. Cyberbullying victims have reported more social difficulties and higher levels of depression and anxiety than victims of traditional bullying (Campbell, Spears, Slee, Butler, & Kift, 2012). Females are more likely than males to be cyber victims (CDC, 2013; Wang et al., 2009). Although research on bullying has burgeoned in recent years, there is little longitudinal research, especially on cyberbullying. In addition, there is minimal research on bullying in rural areas. Given the increased stressors present in rural areas (U.S. Department of Justice, Office for Victims of Crime, 2001), it is vital that researchers gain a better understanding of rural bullying. Much of the existing longitudinal research was conducted outside of the United States and does not distinguish between past, current, and chronic victimization (Barker, Arseneault, Brendgen, Fontaine, & Maughan, 2008; Jose, Kljakovic, Scheib, & Notter, 2012; Lester, Cross, & Shaw, 2012).

LITERATURE REVIEW Bullying in Rural Areas Minimal research has been conducted on health-related risk and protective factors in rural communities (Robbins, Dollard, Armstrong, Kutash, & Vergon, 2008; Witherspoon

Past, Current, and Chronic Bullying Victimization

101

& Ennett, 2011). The impoverished rural area being studied has a constellation of risk factors that likely impact the youth in this sample. Infant mortality is often used as a measure of the health of an area, and the average infant mortality rate of the two counties in this study was 22 per 1,000, 3 times higher than the national average (Heisler, 2012). Further, the unemployment rate was 12%, 5% higher than the national average (Bureau of Labor Statistics, 2012). In addition, the closest large city is 100 miles from both counties, and limited public transportation options make accessing resources present in a city (e.g., a large hospital) problematic. Rural youth are more likely to engage in high-risk behaviors (e.g., substance use, bringing a weapon to school, sexual intercourse) and are at an increased risk for poor educational outcomes, compared to suburban and urban youth (Atav & Spencer, 2002; Witherspoon & Ennett, 2011). Based on these high rates of risk-taking behaviors and unique stressors of rural living (e.g., geographic isolation, minimal community resources; Kusmin, 2008; U.S. Department of Justice, Office for Victims of Crime, 2001), bullying in rural schools might differ from bullying in urban schools. Studies of rural students report that 82% of students reported experiencing some form of victimization (Dulmus et al., 2004) and 33% reported traditional bullying victimization (Price, Chin, Higa-McMillan, Kim, & Frueh, 2013). These prevalence rates are higher than the 19.9% (CDC, 2013) and the 10.6% (Nansel et al., 2001) victimization rates obtained in national studies.

Longitudinal Studies of Victimization Most of the longitudinal studies of bullying victimization have found that victimization is a moderately stable phenomenon (Barker et al., 2008; Jose et al., 2012). Physical/verbal bullying victimization appears to be more stable than cyberbullying victimization (Jose et al., 2012), and both forms of victimization result in enduring negative consequences. One meta-analysis found that childhood bullying victimization led to increased rates of depression that endured an average of 6 years after victimization (Ttofi, Farrington, Losel, & Loeber, 2011). A second meta-analysis found that violent behavior related to childhood victimization persisted an average of 6.9 years following victimization (Ttofi, Farrington, & Losel, 2012). However, these studies have failed to distinguish between past, current, and chronic victimization, making it impossible to determine if duration of victimization affected the severity or prevalence of negative consequences. Longitudinal studies have found that increased victimization leads to low school satisfaction, low levels of perceived social support, and poor mental health outcomes. Haddow (2006) found that “repeated” victimization prior to age 12 years resulted in difficulty concentrating (male victims only) and sleeping (female victims only), low levels of perceived school safety, and increased levels of unhappiness and involvement in school violence. Researchers in England examined students at two time points 6 months apart. At Time 1, victimization had no impact on school satisfaction. However, 6 months later, increased levels of victimization were related to decreased levels of school satisfaction (Boulton, Chau, Whitehand, Amataya, & Murray, 2009). These researchers also showed that children with the highest rates of victimization at Time 1 had the greatest decreases in self-perception 5 months later at Time 2 (Boulton, Smith, & Cowie, 2010). These findings did not distinguish between children with different patterns of victimization (i.e., victimized at Time 1 only, Time 2 only, or at Time 1 and Time 2), making it problematic to draw conclusions about the impact of past, current, or chronic victimization. Using a three-wave longitudinal design and a sample of more than 1,110 American students from 14 schools, Esbensen and Carson (2009) created three groups of children:

102

Smokowski et al.

nonvictims, intermittent victims, and repeat victims. Repeat victims reported increased negative views of school and commitment to negative peers and lower levels of school safety and self-esteem compared to nonvictims and intermittent victims. Using a similar design, Scholte, Engles, Overbeek, de Kemp, and Haselager (2007) conducted a longitudinal study in the Netherlands that compared levels of peer-perceived social behavior in four categories of bullying victims: those who experienced bullying during childhood only, during adolescence only, during both childhood and adolescence, or were nonvictims. Compared with nonvictims, those victimized in both childhood and adolescence had the worst outcomes and were peer rated as being less liked, less cooperative, more shy, and having fewer friends in both childhood and adolescence. Both of these studies highlighted that although episodically and chronically victimized children had more negative perceptions of self and school, the chronically victimized children displayed the worst outcomes.

School Experiences, Social Relationships, and Mental Health of Victimized Adolescents Victimized youth report higher levels of school dissatisfaction and lower rates of school connectedness and school bonding compared to nonvictimized youth (Dulmus, Sowers, & Theriot, 2006; Totura et al., 2008; You et al., 2008). Bullied youth view school as a dangerous place and report higher school disorder (i.e., presence of fighting, problem behavior, and gang involvement) compared to students not involved in bullying (Totura et al., 2008). Perceptions of racial discrimination were positively associated with increased peer nominations for victimization in a sample of African American and Latino youth (Seaton, Neblett, Cole, & Prinstein, 2013), suggesting that racial minorities who are bullied are at an increased risk of perceiving racial discrimination. This assertion was supported in a study of 2,682 Dutch, Turkish, Moroccan, and Surinamese children ages 10–13 years that found that Dutch participants were more likely to report personal victimization, whereas ethnic minorities were more likely to report ethnic discrimination (Verkuyten & Thijs, 2006). Victimized youth perceive lower levels of teacher support (Berkowitz & Benbenishty, 2012; Furlong, Chung, Bates, & Morrison, 1995) and peer support (Demaray & Malecki, 2003; Furlong et al., 1995; Holt & Espelage, 2007) compared to their nonvictimized classmates. Victims of bullying often perceive that teachers and peers are unable and unwilling to stop the bullying, which erodes victims’ sense of support. A 2-year longitudinal study found that in sixth grade, only 17% of peer bystanders intervened in a bullying situation to defend the victim. The rate of supportive bystander behavior increased to only 20% in eighth grade (Salmivalli, Lappalainen, & Lagerspetz, 1998). Youth who are chronically victimized are repeatedly exposed to situations where their peers witness their harassment and fail to help, leaving the victims feeling unsupported and alone. Further, these bullied youth often do not receive support at home and report low levels of maternal support (Holt & Espelage, 2007). This lack of social support is likely to contribute to the poor mental health functioning of victimized adolescents. Indeed, victims of bullying typically report higher rates of depression and anxiety ­compared to bullies, bully-victims, and noninvolved youth (Juvonen, Graham, & Schuster, 2003; Kaltiala-Heino, Rimpela, Marttunen, Rimpela, & Rantanen, 1999; Menesini, Modena, & Tani, 2009). Victims also suffer from low self-esteem (Olweus, 1994), which may cause them to have a negative view of the future. Finally, numerous researchers have found that victims display increased rates of reactive aggression compared to

Past, Current, and Chronic Bullying Victimization

103

n­ onvictimized youth (Camodeca & Goossens, 2005; Camodeca, Goossens, Terwogt, & Schuengel, 2002; Salmivalli & Nieminen, 2002). Based on this past research, we formed the following hypotheses: (a) Physical/verbal bullying and cyberbullying victimization would be a risk factor related to negative developmental outcomes (i.e., negative school experiences, low social support, and poor mental health) in rural adolescents, and (b) chronic physical/verbal bullying and cyberbullying victimization would have more deleterious effects than current or past physical/verbal ­bullying and cyberbullying victimization.

METHOD Participants Participants were enrolled in 28 middle schools in two rural counties within the Southeastern United States. At Time 1 (spring of 2011), participants were in a middle school grade (Grades 6 through 8) and approximately one-third of the sample came from each grade. At Time 2 (Spring of 2012), participants had moved up one grade. Students who moved out of these two school districts were lost to attrition. In County 1, the sample included all middle school students (i.e., a complete census) in public schools. County 2 was much larger than County 1; therefore, a random sample of 40% of public middle school students in County 2 was included in the assessment. Parents from County 2 received a letter explaining the study. If they did not want their child or children to participate, they sent a letter requesting nonparticipation, and their child or children were removed from the study roster. Three parents sent letters of refusal. Students assented to participate by reading and electronically signing an assent screen prior to completing the online assessment. In both counties, students were given the opportunity to decline participation; 60 students declined to participate in the study over the 2 years. This study included only those participants with complete data at both time points and who responded to questions about physical/verbal bullying and cyberbullying victimization (N 5 3,127). The sample was 52.2% female and exceptionally racially diverse: 26.8% identified as American Indian/Native American, 27.3% as White, 24.3% as African American, 8.3% as Hispanic, and 12.1% as mixed race or other. Participants’ mean age was 12.7 years. Two-thirds of participants received free or reduced-price lunch, and 73% lived in families with two adults.

Measures The School Success Profile (SSP; Bowen & Richman, 2008) is a 220-item youth selfreport survey that measures attitudes and perceptions about school, friends, family, neighborhood, self, and health and well-being. This study used the SSP1, which is a modified version of the SSP that includes all original SSP scales in addition to internalizing and externalizing subscales from the Youth Self-Report (i.e., the child form of the Child Behavior Checklist [CBCL]; Achenbach & Ruffle, 2000). A third added scale was a modified version of the Rosenberg (1965) Self-Esteem Scale. Independent Measures. Gender was coded 1 for female and 0 for male. The free or reduced-price school lunch program variable was used as a proxy for socioeconomic status and was coded 1 if the child participated in the program and 0 if he or she did not.

104

Smokowski et al.

Language spoken at home was coded 1 for participants who spoke a language other than English at home and 0 for those who spoke English at home. Age was a continuous variable indicating the participant’s age at study enrollment. Finally, family structure was measured with a dichotomous variable coded 1 for single-parent household and 0 for all other family configurations. Two SSP1 items assessed the respondent’s experience with physical/verbal bullying and cyberbullying victimization: “During the past 12 months, have you ever been bullied on school property?” and “During the past 12 months, have you ever been electronically bullied (including being bullied through e-mail, chat rooms, instant messaging, websites, or texting)?” Both questions used a yes/no response option. These items were identical at Time 1 and Time 2. Following data collection, students were categorized into four groups based on their Time 1 and Time 2 responses to the two victimization questions earlier. Students who reported being a victim of physical/verbal bullying or cyberbullying at both time points were labeled chronic victims, students who reported being victimized at Time 1 but not Time 2 were labeled past victims, students who reported victimization at Time 2 but not Time 1 were labeled current victims, and nonvictims reported no history of being bullied at either time point. Baseline measures of school experiences, mental health, and social support were used in each model to control for Year 1 functioning in predicting Year 2 outcomes. The baseline measures were identical to the Year 2 dependent variables, as described in the following text. Dependent Measures. The dependent measures were indicators of school experiences, social support, and mental health. Three scales measured school experiences: school satisfaction, perceived discrimination, and school hassles. The 7-item school satisfaction scale measured the respondent’s overall satisfaction with school. Example items included, “I enjoy going to this school” and “I get along well with teachers at this school.” Each item was rated on a 3-point Likert scale ranging from not like me to a little like me to a lot like me. Cronbach’s alpha reliability was .85 for this sample. The 3-item perceived discrimination scale assessed how often participants experienced racial discrimination. Example items included, “How often do people dislike you because of your race or ethnicity?” and “How often have you seen friends treated unfairly because of their race or ethnicity?” Each item was rated on a 4-point Likert scale (never, sometimes, frequently, or always); Cronbach’s alpha reliability was .76 in this sample. The 13-item school hassles scale was a measure of unpleasant interactions at school during the past 30 days. Example items included, “Someone treated you in a disrespectful way” and “Someone at school pushed, shoved, or hit you.” The frequency of these events was measured on a 3-point Likert scale (never, once or twice, or more than twice). Cronbach’s alpha reliability for this measure was .92 in this sample. Social support was measured with four subscales that assessed perceived parent, friend, and teacher support and peer rejection. Parent support was assessed using a 5-item scale that measured the frequency of emotional support offered to the respondent during the past 30 days from an adult in the child’s home. Example items included, “How often did the adults in your home let you know that you were loved?” and “How often did the adults in your home tell you that you did a good job?” The frequency of these events was measured on a 3-point Likert scale (never, once or twice, or more than twice); Cronbach’s alpha reliability was .92 in this sample. The 5-item friend support subscale measured the

Past, Current, and Chronic Bullying Victimization

105

student’s perceptions of the extent of support provided by his or her friends. Example items included, “I can count on my friends for support” and “I can trust my friends.” Each item was rated on a 3-point Likert scale (not like me, a little like me, or a lot like me). The Cronbach’s alpha reliability was .91 in this sample. Teacher support was measured using an 8-item subscale that assessed the student’s perception of his or her teachers’ supportive behavior. Example items included, “My teachers care about me” and “My teachers give me a lot of encouragement.” Each item was rated on a 4-point Likert scale (strongly disagree, disagree, agree, or strongly agree); Cronbach’s alpha reliability was .90 in this sample. Peer rejection was measured using a 3-item subscale that assessed student’s perception of peer acceptance. Example items included, “I am made fun of by friends” and “I am picked on by friends.” Each item was rated on a 3-point Likert scale (not like me, a little like me, or a lot like me); Cronbach’s alpha reliability was .72 in this sample. Mental health was assessed using five scales that measured depression, anxiety, externalizing behaviors (i.e., aggression), future optimism, and self-esteem. Achenbach and Ruffle’s (2000) 7-item internalizing subscale from the Youth Self-Report (i.e., child-form CBCL) was divided into a 4-item depression subscale and a 3-item anxiety subscale. Example items from the depression subscale included, “I often feel sad” and “I often feel alone.” Cronbach’s alpha reliability for this scale was .84 in this sample. Example items from the anxiety subscale included, “I often feel nervous or tense” and “I often feel fearful or anxious.” Cronbach’s alpha reliability for this scale was .79 in this sample. Both the depression and anxiety subscales were rated on a 3-point Likert scale (not like me, a little like me, or a lot like me). The 12-item externalizing behaviors scale measured various aggressive and noncompliant behaviors. Example items included, “I get in many fights” and “I break rules at home, school, or elsewhere.” Each item was rated on a 3-point Likert scale (not like me, a little like me, or a lot like me), and Cronbach’s alpha reliability was .87 in this sample. Five items from the Rosenberg Self-Esteem Scale (1965) measured student’s selfesteem. Example items included, “I am able to do things as well as most other people” and “I have confidence in myself.” Each item was assessed on a 3-point Likert scale (not like me, a little like me, or a lot like me). Cronbach’s alpha reliability was .91 in this sample. Future optimism was assessed with 12 items measured on a 4-point Likert scale (strongly disagree, disagree, agree, and strongly agree). Example items included, “I feel positive about the future” and “I make good choices.” Cronbach’s alpha reliability was .94 in this sample. Each of these measures of school experiences, mental health, and social support were assessed during Time 1 and Time 2. Year 1 measures were included in analytic models to control for baseline functioning. For each scale, adding the items and dividing by the number of items answered derived the mean item rating. This strategy reduced missing data.

DATA ANALYSES As described earlier, we cross-classified the two dichotomous questions about p­ hysical/ verbal bullying or cyberbullying victimization at Time 1 and 2. This step yielded a ­physical/verbal victimization variable with four categorical groups: never a victim (n 5 2,157, 69%), past victims (n 5 376, 12%), current victims (n 5 250, 8%), and chronic victims (n 5 344, 11%). Using the same process, the cyberbullying victimization variable

106

Smokowski et al.

had the following groups: never a victim (n 5 2,658, 85%), past victim (n 5 219, 7%), current victim (n 5 156, 5%), and chronic victim (n 5 94, 3%). Considering the 28 schools in our study design, students coming from the same school might share common characteristics on an outcome variable in comparison with students from other schools. Using the intraclass correlation coefficient (ICC) developed by Raudenbush and Bryk (2002), we tested the clustering effects of the outcomes. The results suggested that for most of the scales, less than 2.3% of the variation lies between schools. Teacher support and school satisfaction had ICCs of 4.4% and 5.5%, respectively, which is still low enough to indicate that clustering effects were not present, and a multilevel analysis with an ordinary regression model could safely assume independent observations of the sample data. We proceeded with hierarchical regression with independent variables entered in blocks, yielding five models that predicted each outcome for school experiences, social support, and mental health. The first block included demographic control variables. The second block for past year victimization included two indicators for physical/verbal bullying victimization and cyberbullying victimization during Year 1. The third block added the Year 1 assessment of the dependent variable under consideration, providing a control for baseline functioning and enabling us to evaluate if Year 1 measures of school experiences, mental health, or social support nullified the impact of demographic characteristics or victimization during that baseline year. The fourth block included current physical/verbal bullying and cyberbullying victimization variables. Finally, the fifth block contained one variable indicating chronic physical/verbal victimization and a second variable measuring chronic cyberbullying victimization. Listwise deletion (Allison, 2002) was used to handle missing data. All assumptions for hierarchical multiple regression were met.

RESULTS All three victim groups (i.e., past victims, current victims, and chronic victims) had worse developmental outcomes than nonvictims. Chronic victims had the worst outcomes, and current victims had the next most problematic outcomes. Past victims had poor outcomes on some indicators, but for many outcomes, these direct effects were not statistically significant once the Year 1 dependent variable was entered into the model in the third block.

School Experiences Chronic victimization had pervasive negative effects in predicting lower school satisfaction and higher levels of school hassles and perceived discrimination (see Table 1). Current victimization had effects that were equally widespread and nearly as strong. Physical/verbal bullying victimization effects were stronger than effects for cyberbullying victimization. The direct effects for past victimization, both physical bullying and cyberbullying, were no longer statistically significant after the baseline dependent variable, current victimization, or chronic victimization were entered into the model. This pattern suggests that Year 1 victimization may be associated with negative Year 1 school experiences (lower school satisfaction, stronger perceptions of hassles, higher perceived discrimination), which in turn lead to negative Year 2 school experiences. Past victimization may also influence current or chronic victimization, leading to indirect relationships

Past, Current, and Chronic Bullying Victimization

107

with Year 2 school experiences. These indirect effects are exploratory and should be confirmed in future research.

Social Support Perceptions of parent support in Year 2 were inversely related to past physical/verbal bullying victimization, current cyberbullying victimization, and chronic victimization for both physical/verbal bullying and cyberbullying (see Table 2). Reports of teacher support in Year 2 were inversely related to current physical/verbal bullying and cyberbullying victimization and chronic physical/verbal victimization. Current victims of physical/verbal bullying and cyberbullying and chronic cyberbullying victims reported lower levels of friend support. Past physical/verbal victimization, current physical/verbal bullying and cyberbullying victimization, and chronic physical/verbal and cyber victimization were positively associated with Year 2 peer rejection. TABLE 1.  Episodic and Chronic Victimization and Year 2 School Experiences School Satisfaction

School Hassles

Perceived Discrimination

Demographics   Gender (female)

2.021

.024b

  Free/reduced lunch (yes)

.004

2.009

  Language at home (not English)

.003

.007

 Age

2.052**

2.008

  Single parent family (yes)

2.058***

2.001

.037* .022 .075*** .032a 2.001

Past year victimization   Physical/verbal victim Year 1

2.018a

.000b

  Cyberbully victim Year 1

2.020a

2.025b

.001 .018a

Past year dependent variable   Dependent variable Year 1

.458***

.374***

.349***

  Physical/Verbal victim Year 2

2.082***

.232***

.100***

  Cyberbully victim Year 2

2.020

.094***

.060***

  Physical/Verbal victim Years 1–2

2.088***

.248***

.126***

  Cyberbully victim Years 1–2

2.023

.115***

.100***

.250

.400

.220

85***

163***

69***

Current year victimization

Chronic victimization Years 1 and 2

Adjusted

R2

  F (7; . 2,700) aStatistically bStatistically

significant until Year 1 dependent variable was added to the model. significant until chronic victimization was added to the model. *p , .05. **p , .01. ***p , .001.

108

Smokowski et al.

TABLE 2.  Episodic and Chronic Victimization and Year 2 Social Support Parent Support

Teacher Support

Friend Support

.001

.051**

Peer Rejection

Demographics   Gender (female)

.010

  Free/reduced lunch (yes)

2.032a

  Language at home (not English)

2.023a

 Age

2.067***

  Single parent family (yes)

2.016

.013

2.020a

2.039*

.016

2.013

.007

2.079***

2.049**

.015

2.016

2.007

2.025a

2.053**

2.002

2.014a

.036*

2.005

.015 a

2.022

Past year victimization   Physical/verbal victim Year 1   Cyberbully victim Year 1

2.015

2.023

a

Past year dependent variable   Dependent variable Year 1

.415***

.330***

.406***

.226***

Current year victimization   Physical/verbal victim Year 2

2.016

2.055**

2.054**

.152**

  Cyberbully victim Year 2

2.063***

2.037*

2.034*

.096***

  Physical/verbal victim Years 1–2

2.047**

2.081***

2.014

.164***

  Cyberbully Victim Years 1–2

2.034*

2.001

2.036*

.055**

.20

.13

.19

.14

38***

98***

40***

Chronic victimization Years 1 and 2

Adjusted

R2

  F (7; . 2,700)

62***

aStatistically

significant until Year 1 dependent variable was added to the model. *p , .05.**p , .01.***p , .001.

Mental Health Past physical/verbal bullying victimization, current physical/verbal bullying and cyberbullying victimization, and chronic physical/verbal victimization were all associated with lower levels of future optimism in Year 2, controlling for all other factors in the model (see Table 3). Self-esteem in Year 2 was inversely related to current physical/verbal bullying and cyberbullying victimization and chronic physical/verbal bullying and cyberbullying victimization. The same pattern was evident for anxiety and aggression in Year 2, except chronic physical/verbal victimization was not significantly related to Year 2 aggressive behavior.

DISCUSSION There is no doubt that physical/verbal bullying and cyberbullying are a serious concern for a significant number of youth in the United States. In this large, ethnically diverse

Future Optimism

Self-Esteem

Depression

Anxiety

Aggression

0.109***

0.116***

0.050**

0.029

0.017a

0.020a

0.015a

Demographics   Gender (female)

0.066***

20.034*

  Free/reduced lunch (yes)

20.031

  Language at home (not English)

20.031a

20.007

0.030a

0.017

0.009

 Age

20.091***

20.067***

0.032**

0.072***

0.026a

  Single parent family (yes)

20.011

0.030

0.036*

0.028a

20.027

0.024a

0.031a

0.016

0.007

0.010

20.017

20.008

20.028

0.317***

0.368***

0.006

Past year victimization   Physical/verbal victim Year 1   Cyberbully victim Year 1

20.018**

Past, Current, and Chronic Bullying Victimization

TABLE 3.  Episodic and Chronic Victimization and Year 2 Mental Health

Past year dependent variable   Dependent variable Year 1

0.433***

0.351***

0.482***

Current year victimization   Physical/verbal victim Year 2

20.037*

20.082**

0.048**

0.042*

0.056**

  Cyberbully victim year 2

20.047**

20.105***

0.100***

0.089***

0.073***

  Physical/verbal victim Years 1–2

20.039*

20.088***

0.078***

0.052**

0.025

  Cyberbully victim Years 1–2

20.011

20.060***

0.089***

0.086***

0.055**

Chronic victimization Years 1 and 2

Adjusted

R2

  F (7; . 2,700)

0.19

0.28

0.20

0.28

36***

56***

91***

57***

89***

significant until Year 1 dependent variable was added to the model. *p , .05.**p , .01.*** p , .001.

109

aStatistically

0.13

110

Smokowski et al.

sample of 3,127 rural youth, 35% of middle school students reported being victimized in some way over the course of 2 years. For physical/verbal bullying, 12% had been victimized in Year 1, 8% in Year 2, and another 11% in both Years 1 and 2. Despite being in an impoverished, rural context where cell phone coverage is sporadic and computers are not commonly available in homes, cyberbullying was less prevalent but still widespread. For cyberbullying, 7% of students had been victimized in Year 1, 5% in Year 2, and another 3% in both Years 1 and 2. Our victimization prevalence rates are reasonable estimates, particularly considering the dearth of research on rural youth. Rates from this study fall between the high of 82.3% of rural students who reported experiencing some form of victimization by Dulmus and colleagues (2004) and the nationally representative sample prevalence rates of 19.9% victimization found in the CDC (2013) data or the 10.6% victimization rate reported by Nansel and colleagues (2001). Because this study used the same victimization items as the CDC, there is clear evidence that victimization rates are elevated in the rural areas we studied. While being cautious in generalizing this information to other rural areas, concern about bullying victimization in rural areas is warranted, as is the necessity of conducting more research with rural youth. The results from this study underscore the importance of conducting longitudinal studies on victimization. Timing and chronicity of victimization experiences were critical factors to study in relating bullying to developmental outcomes. Current, past, and chronic physical/verbal bullying and cyberbullying victimization were related to lower levels of school satisfaction, perceived social support, and mental health. This evidence suggests that any amount of bullying victimization, even discrete instances that do not endure into the following school year, have serious, deleterious effects for children. Both types of bullying victimization had widespread impacts across developmental domains, hampering academic experiences, social interactions, and mental health processes. This evidence strengthens the case for both types of bullying victimization to be considered interpersonal traumas that precipitate feelings of shame and humiliation, leading to profound damage to self-identity and interpersonal functioning. Feelings of shame and humiliation that can derive from bullying victimization are likely to result in an impaired ability to process emotions, an eroded experience of the self and later psychopathology (Lee, Scragg, & Turner, 2001). Our hypotheses were supported. Past, current, and chronic physical/verbal bullying and cyberbullying victimization were risk factors related to pervasive negative developmental outcomes in children. Chronic victimization displayed the worst effects across all developmental outcomes, which is in line with the notion that cumulative risk factors create a “pileup” that is more detrimental than individual risk factors (Davies, 1999). Chronic victimization was associated with increased perceptions of school hassles, racial discrimination, peer rejection, depression, anxiety, and aggression with concomitant decreases in school satisfaction, future optimism, self-esteem, and support from parents, teachers, and friends. These findings confirm past research (Esbensen & Carson, 2009) and extend the discussion to impoverished rural settings. Bullying victims should be encouraged to seek help so that victimization does not become chronic. Current victimization was nearly as deleterious as chronic victimization and was more closely tied to Year 2 developmental outcomes than past year victimization. Consequently, it is paramount for school personnel to intervene in current bullying dynamics whether or not the situation has been going on long term. Current bullying, either physical/verbal or cyber, should not be tolerated. Adults should not underestimate the significance of the

Past, Current, and Chronic Bullying Victimization

111

isolated, current event, while waiting to see whether the victimization becomes chronic. Current bullying victimization has a pervasive, negative effect on child functioning that needs to be addressed when it happens. It was an encouraging sign that victimization in the past year did not have statistically significant lingering effects on developmental outcomes in the current year. Perhaps some of the trauma resulting from past bullying victimization heals with time. The diminishing effect of past victimization appeared to be the case with cyberbullying victimization in particular. At the same time, past physical/verbal bullying victimization had such a traumatic effect that it displayed a continuing relationship with perceptions of low parent support, low future optimism, and high peer rejection 1 year after the victimization. Past cyberbullying victimization also continued to impact students’ perceptions of school danger in the current year. The persistence of these effects underscores the importance of having more longitudinal research on bullying. Most of the direct effects of past episodes of victimization lost their statistical significance when Year 1 measures of school satisfaction, social support, and mental health or Year 2 victimization variables were added to the models. This pattern suggests that Year 1 victimization may be associated with negative Year 1 school experiences, lack of Year 1 social support, or Year 1 mental health problems, which in turn lead to negative Year 2 outcomes. Consequently, the direct impact of past victimization might fade, but the damage to well-being may continue indirectly by shifting victims onto a problematic trajectory that persists over time. These indirect effects are exploratory and should be confirmed in future research that uses sophisticated analytic techniques to examine influential pathways. The effects for physical/verbal bullying victimization were usually slightly stronger than those for cyberbullying victimization. It was intuitive that physical/verbal bullying would erode feelings of school safety and satisfaction because the bullying occurs in the school environment. However, it was surprising that cyberbullying also had significant negative impacts on school experiences. Although cyberbullying often occurs outside of school, the detrimental effects clearly impacted children’s ability to enjoy school and feel safe during the school day. Given that cyberbullies often conceal their identities, victims may feel constantly unsettled (Hoff & Mitchell, 2009), which might explain why cyber victimization resulted in an increased perception of danger and lower levels of school satisfaction. The anonymity of cyberbullying might prime victims to see danger everywhere, especially at school, and may result in a state of hypervigilance or heightened sensitivity to threats. Hypervigilance might cause victims to perceive discrimination, school hassles, and school danger more frequently. Levels of perceived social support varied by type of victimization (i.e., physical/verbal bullying or cyberbullying) and by source of support. Because physical/verbal bullying occurred at school, it is likely that victims’ friends witnessed the bullying and might not have assisted the victim, which could account for the low levels of friend support reported by current victims of physical/verbal bullying. It is well documented that victims of physical/verbal bullying have poor peer relationships (Nansel et al., 2004; Nansel et al., 2001). Unlike physical victimization in which the bully might be stronger than the victim and friends, in situations of cyberbullying, friends can intervene by posting supportive messages on social networks. A lack of cyber support from friends might deeply wound the victim because this is about courage not strength. It is noteworthy that all three victim groups of both physical/verbal bullying and cyberbullying reported lower levels of parent and teacher support than nonvictims. Either parents and teachers are not intervening to help stop bullying or victims are not telling their

112

Smokowski et al.

parents and teachers about the bullying leaves these adults with a lack of knowledge. Both scenarios would result in the victims’ decreased perceptions of adult support. Future interventions should focus on empowering victims to report the bullying to a parent, teacher, or trusted adult. The simple act of reporting victimization is the first step to helping victims create support networks. For mental health outcomes, all current and chronic victims of both physical/verbal bullying and cyberbullying reported higher levels of anxiety, depression, and aggressive behaviors as well as lower levels of self-esteem and future optimism than nonvictims. Perhaps the shame and humiliation caused by victimization is a partial explanation for the poor mental health outcomes of victims. Shame causes self-blame and a negative view of self (Lee et al., 2001), which might translate into internalizing disorders such as depression and anxiety. The experience of humiliation, which fosters a negative view of and anger toward the perpetrator (Lee et al., 2001), might explain the higher levels of aggressive behaviors among victims. Future research should consider the role of shame and humiliation in the relationship between bullying victimization and negative mental health outcomes. The widespread effects of chronic victimization on many developmental outcomes should alarm school staff, especially guidance counselors and social workers. We extended past research by showing that repeated victimization dramatically erodes mental health over time. Special consideration should be given to the needs of adolescent females. Relative to males, females reported more depression, anxiety, aggression, perceived discrimination, and lower self-esteem. These results for females are in line with previous studies and national data showing females to be at high risk for certain forms of bullying victimization (e.g., cyber, relational/social, verbal) and mental health difficulties (CDC, 2013; Hankin, 2006; Negriff & Susman, 2011; Wang et al., 2009). Cyberbullying, in particular, is nearly twice as common among females as it is among males (22% vs. 11% respectively; CDC, 2013). Given this high level of risk for females, school counselors and social workers should invest in gender-specific interventions that foster social support and address mental health issues.

Limitations First, using a dichotomous variable to measure bullying victimization resulted in a loss of variability as a child who had been victimized once was treated synonymously with a child who was victimized multiple times in each year. Second, it is possible that using a single item to assess the prevalence rates of physical/verbal bullying and cyberbullying victimization in this sample resulted in an underestimation of true victims. Researchers have found that using multiple items to assess bullying victimization yielded a higher count of victims than using only one item (Esbensen & Carson, 2009). However, because of limited student time and space on the assessment instrument, adding additional items was not feasible in this study. Third, a definition of bullying was not provided to students when completing the assessment, which might have made it difficult for students to identify situations in their own lives that constituted bullying.

CONCLUSION Any amount of bullying victimization can result in negative outcomes. However, chronic victimization is clearly more detrimental than past or current victimization. Ongoing victimization may serve as a form of interpersonal trauma that influences school e­ xperiences,

Past, Current, and Chronic Bullying Victimization

113

personal relationships, and mental health functioning. Victims of chronic bullying are clearly in need of additional supports. Ideally, school personnel should intervene in bullying dynamics before a child becomes a chronic victim. However, if this intervention does not occur, chronic victims should be provided with ample supports and appropriate mental health treatment.

REFERENCES Achenbach, T. M., & Ruffle, T. M. (2000). The child behavior checklist and related forms for assessing behavioral/emotional problems and competencies. Pediatrics in Review, 21(2), 265–271. Allison, P. (2002). Missing data. Thousand Oaks, CA: Sage. Atav, S., & Spencer, G. A. (2002). Health risk behaviors among adolescents attending rural, suburban, and urban schools: A comparative study. Family and Community Health, 17(12), 53–64. http://dx.doi.org/10.1097/00003727-200207000-00007 Barker, E. D., Arseneault, L., Brendgen, M., Fontaine, N., & Maughan, B. (2008). Joint development of bullying and victimization in adolescence: Relations to delinquency and self-harm. American Academy of Child and Adolescent Psychiatry, 47, 1030–1038. http://dx.doi.org/10.1097/ CHI.0b013e31817eec98 Berkowitz, R., & Benbenishty, R. (2012). Perceptions of teachers’ support, safety, and absence from school because of fear among victims, bullies, and bully-victims. American Journal of Orthopsychiatry, 82(67), 67–74. http://dx.doi.org/10.111/j.1939-0025.2011.01132.x Boulton, M. J., Chau, C., Whitehand, C., Amataya, K., & Murray, L. (2009). Concurrent and shortterm longitudinal associations between peer victimization and school and recess liking during middle childhood. British Journal of Educational Psychology, 79, 207–221. http://dx.doi .org/10.1348/000709908X336131 Boulton, M. J., Smith, P. K., & Cowie, H. (2010). Short-term longitudinal relationships between children’s peer victimization/bullying experiences and self-perceptions: Evidence for reciprocity. School Psychology International, 31, 296–311. http://dx.doi.org/10.1177/0143034310362329 Bowen, G. L., & Richman, J. M. (2008). The School Success Profile. Chapel Hill, NC: University of North Carolina at Chapel Hill. Bureau of Labor Statistics. (2012). Local area unemployment statistics map. Retrieved from http:// www.bls.gov/lau/ Camodeca, M., & Goossens, F. A. (2005). Aggression, social cognitions, anger, and sadness in bullies and victims. Journal of Child Psychology and Psychiatry, 46(2), 186–197. Camodeca, M., Goossens, F. A., Terwogt, M. M., & Schuengel, C. (2002). Bullying and victimization among school-aged children: Stability and links to proactive and reactive aggression. Social Development, 11(3), 332–345. Campbell, M., Spears, B., Slee, P., Butler, D., & Kift, S. (2012). Victims’ perceptions of traditional and cyber bullying, and the psychosocial correlates of their victimization. Emotional & Behavioural Difficulties, 17(3–4), 389–401. http://dx.doi.org/10.1080/13632752.2012.704316 Centers for Disease Control and Prevention. (2013). 1991-2011 High School Youth Risk Behavior Survey data. Retrieved from http://apps.nccd.cdc.gov/youthonline Cho, J. I., Hendrickson, J. M., & Mock, D. R. (2009). Bullying status and behavior patterns of preadolescents and adolescents with behavioral disorders. Education & Treatment of Children, 32(4), 655–671. http://dx.doi.org/10.1353/etc.0.0080 Davies, B. (1999). Child development: A practitioner’s guide. New York, NY: Guilford Press. Demaray, M. K., & Malecki, C. K. (2003). Perceptions of the frequency and importance of social support by students classified as victims, bullies, and bully/victims in an urban middle school. School Psychology Review, 32(3), 471–489. Dulmus, C. N., Sowers, K. M., & Theriot, M. T. (2006). Prevalence and bullying experiences of victims and victims who become bullies (bully-victims) at rural schools. Victims and Offenders, 1(15), 15–31.

114

Smokowski et al.

Dulmus, C. N., Theriot, M. T., Sowers, K. M., & Blackburn, J. A. (2004). Student reports of peer bullying victimization in a rural school. Stress, Trauma, and Crisis, 7, 1–16. http://dx.doi .org/10.1080/15434610490281093 Esbensen, F., & Carson, D. C. (2009). Consequences of being bullied: Results from a longitudinal assessment of bullying victimization in a multisite sample of American students. Youth and Society, 41(2), 209–232. http://dx.doi.org/10.1177/0044118X09351067 Frisen, A., Holmqvist, K., & Oscarsson, D. (2008). 13-year-olds’ perception of bullying: Definitions, reasons for victimization, and experience of adults’ response. Educational Studies, 34, 105–117. http://dx.doi.org/10.1080/03055690701811149 Furlong, M. J., Chung, A., Bates, M., & Morrison, R. L. (1995). Perceptions of the frequency and importance of social support by students classified as victims, bullies, and bully/victims in an urban middle school. School Psychology Review, 32(3), 282–298. Gastic, B. (2008). School truancy and the disciplinary problems of bullying victims. Educational Review, 60, 391–404. http://dx.doi.org/10.1080/00131910802393423 Gini, G. (2008). Associations between bullying behavior, psychosomatic complaints, emotional, and behavioral problems. Journal of Pediatrics and Child Health, 44, 492–497. http://dx.doi .org/10.1111/j.1440-1754.2007.01155.x Guerra, N. G., Williams, K. R., & Sadek, S. (2011). Understanding bullying and victimization during childhood and adolescence: A mixed methods study. Child Development, 82, 295–310. http://dx.doi.org/10.1111/j.1467-8624.2010.01556.x Haddow, J. L. (2006). Residual effects of repeated bully victimization before the age of 12 on adolescent functioning. Journal of School Violence, 5(2), 37–52. http://dx.doi.org/10.1300/ J202v05n02_04 Hankin, B. L. (2006). Adolescent depression: Description, causes, and interventions. Epilepsy & Behavior, 8, 102–114. http://dx.doi.org/10.1016/j.yebeh.2005.10.012 Heisler, E. J. (2012). The U.S. infant mortality rate: International comparisons, underlying factors, federal programs. Congressional Research Service. Retrieved from http://www.fas.org/sgp/crs/ misc/R41378.pdf Hinduja, S., & Patchin, J. (2007). Overview of cyber bullying. Retrieved from http://www.stopbullying.gov/at-risk/groups/lgbt/white_house_conference_materials.pdf Hoff, D. L., & Mitchell, S. N. (2009). Cyber bullying: Causes, effects, and remedies. Journal of Educational Administration, 47(5), 652–665. http://dx.doi.org/10.1108/09578230910981107 Holt, M. K., & Espelage, D. L. (2007). Perceived social support among bullies, victims, and bullyvictims. Journal of Youth and Adolescence, 36, 984–994. Jose, P. E., Kljakovic, M., Scheib, E., & Notter, O. (2012). The joint development of traditional bullying and victimization with cyber bullying and victimization in adolescence. Journal of Research on Adolescence, 22(2), 301–309. http://dx.doi.org/10.1111/j.1532-7795.2011.00764.x Juvonen, J., Graham, S., & Schuster, M. A. (2003). Bullying among young adolescents: The strong, the weak, and the troubled. Pediatrics, 112, 1231–1237. Retrieved from http://pediatrics .aappublications.org/content/112/6/1231.full.html Kaltiala-Heino, R., Rimpela, M., Marttunen, M., Rimpela, A., & Rantanen, P. (1999). Bullying, depression, and suicidal ideation in Finnish adolescents: School survey. British Journal of Medicine, 319, 348–351. http://dx.doi.org/10.1136/bmj.319.7206.348 Kusmin, L. D. (2008). Rural America at a glance, 2008 edition (USDA Economic Research Service, Pub. No. EIB-40). Retrieved from http://www.ers.usda.gov/Publications/EIB40/EIB40.htm Lee, D. A., Scragg, P., & Turner, S. (2001). The role of shame and guilt in traumatic events: A clinical model of shame-based and guilt-based PTSD. British Journal of Medical Psychology, 74, 451–466. http://dx.doi.org/10.1348/000711201161109 Lester, L., Cross, D., & Shaw, T. (2012). Problem behaviors, traditional bullying and cyber bullying among adolescents: Longitudinal analysis. Emotional & Behavioural Difficulties, 17, 435–447. http://dx.doi.org/10.1080/13632752.2012.704313

Past, Current, and Chronic Bullying Victimization

115

Menesini, E., Modena, M., & Tani, F. (2009). Bullying and victimization in adolescence: Concurrent and stable roles and psychological health symptoms. The Journal of Genetic Psychology, 170(2), 115–133. Nansel, T. R., Craig, W., Overpeck, M. D., Saluja, G., & Ruan, W. J. (2004). Cross-national consistency in the relationship between bullying behaviors and psychosocial adjustment. Archives of Pediatric and Adolescent Medicine, 158, 730–736. http://dx.doi.org/10.1001/arch pedi.158.8.730 Nansel, T. R., Overpeck, M., Pilla, R. S., Ruan, W. J., 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. http://dx.doi.org/10.1001/ jama.285.16.2094 Nation, M., Vieno, A., Perkins, D., & Santinello, M. (2008). Bullying in school and adolescent sense of empowerment: An analysis of relationships with parents, friends, and teachers. Journal of Community & Applied Social Psychology, 18(3), 211–232. http://dx.doi.org/10.1002/casp.921 National Center for Educational Statistics, U.S. Department of Education. (2011). Student reports of bullying and cyber-bullying: Results from the 2009 School Crime Supplement to the National Crime Victimization Survey. (Web Tables: NCES 2011–336). Retrieved from http://nces. Ed.gov/pubs2011/2011336.pdf Negriff, S., & Susman, E. J. (2011). Pubertal timing, depression, and externalizing problems: A framework, review, and examination of gender differences. Journal of Research on Adolescence, 21(3), 717–746. http://dx.doi.org/10.1111/j.1532-7795.2010.00708.x Olweus, D. (1993). Bullying at school: What we know and what we can do. Malden, MA: Blackwell. Olweus, D. 1994. “Annotation: Bullying at school: Basic facts and effects of a school based intervention program.” Journal of Child Psychology and Psychiatry, 35, 1171–1190. http://dx.doi. org/10.1111/j.1469-7610.1994.tb01229.x Price, M., Chin, M. A., Higa-McMillan, C., Kim, S., & Frueh, B. C. (2013). Prevalence and internalizing problems of ethnoracially diverse victims of traditional and cyber bullying. School Mental Health, http://dx.doi.org/10.1007/s12310-013-9104-6 Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and data analysis methods (2nd ed.). Thousand Oaks, CA: Sage. Robbins, V., Dollard, N., Armstrong, B. J., Kutash, K., & Vergon, K. S. (2008). Mental health needs of poor suburban and rural children and their families. Journal of Loss and Trauma, 13, 94–122. http://dx.doi.org/10.1080/15325020701769170 Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University. Salmivalli, C., Lappalainen, M., & Lagerspetz, K. M. J. (1998). Stability of change behavior in connection to bullying in schools: A two-year follow-up. Aggressive Behavior, 24, 205–218. Retrieved from http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)1098-2337 Salmivalli, C., & Nieminen, E. (2002). Proactive and reactive aggression among school bullies, victims, and bully-victims. Aggressive Behavior, 28, 30–44. Scholte, R., Engles, R., Overbeek, G., de Kemp, R., & Haselager, G. (2007). Stability in bullying and victimization and its association with social adjustment in childhood and adolescence. Journal of Abnormal Child Psychology, 35, 217–228. http://dx.doi.org/10.1007/s10802-0069074-3 Seaton, E. K., Neblett, E. W., Cole, D. J., & Prinstein, M. J. (2013). Perceived discrimination and peer victimization among African American and Latino youth. Journal of Youth and Adolescence, 42, 342–350. http://dx.doi.org/1007/s10964-012-9848-6 Smokowski, P. R., & Holland, K. (2005). Bullying in school: Correlates, consequences, and intervention strategies for school social workers. Children & Schools, 27(2), 101–110. Sweeting, H., Young, R., West, P., & Der, G. (2006). Peer victimization and depression in early-mid adolescence: A longitudinal study. British Journal of Educational Psychology, 76, 577–594. http://dx.doi.org/10.1348/000709905X49890

116

Smokowski et al.

Totura, C. M. W., Mackinnon-Lewis, C., Gesten, E. L., Gadd, R., Divine, K. P., Dunham, S., & Kamboukos, D. (2008). Bullying and victimization among boys and girls in middle school: The influence of perceived family and school contexts. The Journal of Early Adolescence, 29, 571–609. http://dx.doi.org/10.1177/0272431608324190 Tsang, S. K. M., Hui, E. K. P., & Law, B. C. W. (2012). Bystander position taking in school bullying: The role of positive identity, self-efficacy, and self-determination. International Journal of Child Health and Human Development, 5, 103–110. Retrieved from https://www.novapublishers .com/catalog/product_info.php?products_id=6196 Ttofi, M. M., Farrington, D. P., & Losel, F. (2012). School bullying as predictor of violence later in life: A systematic review and meta-analysis of prospective longitudinal studies. Aggression and Violent Behavior, 17, 405–418. http://dx.doi.org/10.1016/j.avb.2012.05.002 Ttofi, M. M., Farrington, D. P., Losel, F., & Loeber, R. (2011). Do 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(11), 63–73. http://dx.doi .org/10.1108/17596591111132873 U.S. Department of Justice, Office for Victims of Crime. (2001). Rural victim assistance: A victim/ witness guide for rural prosecutors (NCJ Pub. No. 211106). Retrieved from http://www.ojp .usdoj.gov/ovc/publications/infores/rural_victim_assistance/ Verkuyten, M., & Thijs, J. (2006). Ethnic discrimination and global self-worth in early adolescents: The mediating role of ethnic self-esteem. International Journal of Behavioral Development, 30(2), 107–116. http://dx.doi.org/10.1177/0165025406063573 Wang, J., Iannotti, R. J., & Nansel, T. R. (2009). School bullying among adolescents in the United States: Physical, verbal, relational, and cyber. Journal of Adolescent Health, 45, 368–375. http://dx.doi.org/10.1016/j.jadohealth.2009.03.021 Witherspoon, D., & Ennett, S. (2011). Stability and change in rural youths’ educational outcomes through the middle and high school years. Journal of Youth and Adolescence, 40(9), 1077–1090. http://dx.doi.org/10.1007/s10964-010-9614-6 You, S., Furlong, M. J., Felix, E., Sharkey, J. D., Tanigawa, D., & Green, J. G. (2008). Relations among school connectedness, hope, life satisfaction, and bully victimization. Psychology in the Schools, 45(5), 446–460. Acknowledgments. This study was funded through a cooperative agreement with the U.S. Centers for Disease Control and Prevention’s National Center for Injury Prevention and Control (5 U01 CE001948-03). Correspondence regarding this article should be directed to Paul R. Smokowski, PhD, University of North Carolina at Chapel Hill, School of Social Work, CB # 3550, 325 Pittsboro Street, Chapel Hill, NC 27599-3550. E-mail: [email protected]

Violence and Victims, Volume 27, Number 3, 2012

Teachers Bullied by Students: Forms of Bullying and Perpetrator Characteristics Teemu Kauppi, MA Maili Pörhölä, PhD University of Jyväskylä, Finland The focus of this study is on the forms in which the bullying of school teachers by students manifests itself, the characteristics of the students who engage in the bullying, and the manner in which the students who engage in bullying behave in their own peer relationships. The data was gathered from primary and lower secondary school teachers by means of an Internet survey. The answers of 70 teachers who had experienced bullying by their students are examined. The teachers had been exposed to different forms of bullying by students. They had typically been bullied by male students. In most cases, the bullying had been perpetrated by an individual student or a small group of students. According to the teachers’ assessment, the majority of the students who bullied them also bullied their fellow students.

Keywords: bullying; harassment; students; teachers; violence against teachers

A

school is an institution where teachers and students work in cooperation to reach the educational objectives set for the students. The most important professional duties of a teacher include not only seeing to it that these educational objectives are met but also assessing the students’ performance, maintaining order, and taking care of the well-being of the students. In addition to the well-being of students, the well-being of teachers has a central role in any school community. It can be assumed that teachers who feel comfortable in their position and are content with their working conditions have a better chance to succeed in supporting the work of their students. Correspondingly, teachers who are not comfortable in their work or lack a feeling of well-being may not be able to perform to their own satisfaction in the demanding position. The experience of being subjected to bullying at work forms a major threat to a teachers’ well-being. It has been discovered that for teachers, the four main sources of bullying are students (e.g., De Wet, 2010a; James et al., 2008; Terry, 1998), colleagues (e.g., Cemaloglu, 2007; Van Dick & Wagner, 2001), superiors (e.g., Blase & Blase, 2003; De Wet, 2010b; Van Dick & Wagner, 2001), and the parents of students (e.g., Benefield, 2004; Fisher & Kettl, 2003). One of the special features of teacher’s work is that teachers can be subjected to bullying by people whose status within the institution is lower than theirs—that is, by students. This type of bullying relationship is quite special in nature.

© 2012 Springer Publishing Company

117

118

Kauppi and Pörhölä

Researchers have used various terms to describe mental and physical violence directed at teachers by their students. For example, terms such as bullying (De Wet, 2010a; James et al., 2008; Terry, 1998), harassment (Kauppi & Pörhölä, 2009), victimization (e.g., Dworkin, Haney, & Telschow, 1988), and violence against teachers (Chen & Astor, 2009; Dzuka & Dalbert, 2007; Khoury-Kassabri, Astor, & Benbenishty, 2009) have been applied to cover this type of violence. In this article, we use the term bullying as an umbrella term to describe both mental and physical violence directed at teachers by their students. The experience of being subjected to bullying at work is known to have a considerably detrimental effect on victims’ health and well-being (Björkqvist, Österman, & Hjelt-Bäck, 1994; Hoel, Faragher, & Cooper, 2004; Leymann & Gustafsson, 1996; Matthiesen & Einarsen, 2004). It has been further discovered that bullying and violence have negative effects on the quality of teachers’ work performance (De Wet, 2010a; Fisher & Kettl, 2003). Earlier studies (e.g., Dzuka & Dalbert, 2007; James et al., 2008; Khoury-Kassabri et al., 2009; Terry, 1998) have already provided some information on the prevalence of victimization of school teachers by students and on the most typical forms in which such victimization manifests itself. Both students’ (e.g., Chen & Astor, 2009; James et al., 2008; Khoury-Kassabri et al., 2009) and teachers’ (e.g., De Wet, 2010a; Dzuka & Dalbert, 2007; Terry, 1998) reports have been examined. However, little is known about the special characteristics of this type of bullying (see De Wet, 2010a; Dzuka & Dalbert, 2007; Kauppi & Pörhölä, 2009). In this article, we describe and analyze the forms of direct and indirect bullying employed on teachers by students and look at what types of students engage in bullying teachers.

DEFINITION OF BULLYING Most studies of school and workplace bullying define bullying with the help of the following three criteria: (a) bullying is when someone directs aggressive behavior towards another party or intentionally hurts and harms another party, (b) bullying manifests repeatedly over a lengthy period of time, and (c) there exists such an imbalance of power between the party perpetrating the bullying and the party being subjected to bullying that the latter cannot defend himself or herself (see Pörhölä, Karhunen, & Rainivaara, 2006, for a concept analysis). The bullying of teachers by students differs significantly in nature from school and workplace bullying taking place at peer level. When a student bullies his or her teacher, there exists a situation where a party holding an inferior status position within the institution bullies a party holding a superior status position. This can be considered to constitute so-called “cross-peer abuse” (Terry, 1998). The teacher experiencing bullying has, at least to begin with, power over the student on the grounds of his or her position as a teacher (e.g., Chan, 2009). In Finland, for example, a teacher is entitled to levy punishment on students for behavior that breaks school rules or norms by giving them detention or removing them from the classroom for the remainder of the class period (Basic Education Act 2003/477, 7:36§, 36a§, 36b§). At the same time, a teacher has the responsibility for the assessment of the students’ learning results and processes. Consequently, the criterion concerning the imbalance of the power relationship cannot be applied in the definition of the bullying of teachers by students in the same manner as it is applied in the definition of bullying taking place at peer level.

Teachers Bullied by Students

119

However, research literature presents only a few definitions for the bullying of teachers by students. For example, according to a definition by Terry (1998, p. 261), [b]ullying occurs in situations where the victim cannot easily escape. It occurs when an uneven balance of power is exploited and abused by an individual or individuals who in that particular circumstance have advantage. Bullying is characterized by persistent, repetitive acts of physical or psychological aggression.

This definition takes into account the imbalance of the power relationship between the parties in a certain situation. Hence, despite teacher’s position as an institutional authority at school, in a particular circumstance, students could achieve enough power over the teacher to enable bullying. In their definition of workplace bullying, Einarsen, Hoel, Zapf, and Cooper (2003) emphasize that, in bullying processes, power can be reversed over the course of bullying or as a consequence of it, so that the victim ends up in an inferior position and has difficulties in defending himself or herself. In the definition by Terry (1998), this development is seen possible also in the relationships between teachers and students. Further, as characterized by Twemlow, Fonagy, Sacco, and Brethour (2006, p. 191), a bullying student is “a student who tends to control the classroom with disruptive behavior that implies contempt for the teacher and who uses coercive tactics to deskill the teacher.” Correspondingly, Dzuka and Dalbert (2007, p. 253) defined violence against teachers by students as “aggressive behavior intended to harm the teacher, which students perpetrate repeatedly and intentionally over a certain amount of time.” With the exception of the imbalance of the power relationship, this definition incorporates the general criteria of bullying presented earlier (see Pörhölä et al., 2006). Like most definitions of school bullying, the one given by Dzuka and Dalbert is clearly perpetrator-oriented. Bullying is defined through the actions of the bully. In the literature on workplace bullying, definitions are not perpetraror-oriented to the same extent; common criteria for bullying include the victim’s experience of being bullied and perceived damage caused to the victim by the bullying (see Pörhölä et al., 2006; Rayner & Keashly, 2005). As far as definitions of bullying are concerned, the criterion that differentiates bullying from, for example, aggression or conflict—both of which can occur as single incidents—is the recurrence of communication that causes harm to the other party (Keashly & Nowell, 2003). However, it would be shortsighted to categorically define the bullying of teachers by students as a recurring series of events from the side of the particular bully or bullies. It is possible, for example, that recurrent acts of threatening, physical violence, insulting verbal comments, or publication of defamatory writings on the Internet, even if perpetrated on a single-time basis by various individual students, can be perceived by a teacher as bullying. Further, as Smith (2012) noted, when bullying occurs in technologically mediated forms, the use of repetition as a criterion for bullying is rather problematic. The act of cyberbullying may repeat itself without the contribution of the bully. For example, if insulting content is uploaded onto a web page, every hit on that page could count as repetition (Smith, 2012). In such a situation, a teacher’s experience is that he or she is the target of recurrent bullying. Another aspect that has often been included in definitions of bullying is the intentionality of the communication that causes harm to the other party. When the focus is on the examination of the experience of the bullied person, however, the intentionality of the bully’s actions is not considered to be a defining factor in the bullying experience. Rayner, Hoel, and Cooper (2002), for example, recommend that no definition of bullying should rest on the aspect of intentionality. According to them, the experience of the victim of the

120

Kauppi and Pörhölä

bullying should not be invalidated on the grounds that the party perceived as the bully has not acted or states that he or she has not acted with the intent to harm the other party. The party at whom the communication is directed may feel bullied irrespective of the intentions of the offending party. In such a case, an approach focusing on the experiencer should be adopted, and the phenomenon should be examined as a bullying experience. Because, in this study, we look at bullying through the experiences of teachers, our definition of bullying does not consider whether or not the perpetrator(s) intentionally engage(s) in bullying. For the purposes of this study, we define the bullying of teachers by students as a communication process in which “a teacher is repeatedly subjected, by one or more students, to interaction that he or she perceives as insulting, upsetting, or intimidating. Bullying can be verbal, nonverbal, or physical in nature.”

THE FORMS OF BULLYING EXPERIENCED BY TEACHERS The bullying of teachers by students usually manifests as insulting, hostile, and unethical verbal and nonverbal communication,for example, by means of name-calling, use of inappropriate language, use of insulting gestures, refusal to cooperate, intimidating or upsetting behavior, sexual harassment, physical violence, and destroying teachers’ property (Chen & Astor, 2009; De Wet, 2010a; Dzuka & Dalbert, 2007; James et al., 2008; Kivivuori & Tuominen, 1999; Lahelma, Palmu, & Gordon, 2000; Terry, 1998). Hostile and unethical communication is, in fact, considered to form one central element of mental violence occurring both in the context of school and working life (Leymann, 1996; Pörhölä et al., 2006). Teachers feel that they are also subjected to indirect forms of bullying by their students. This is true even though the indirect forms of bullying used by students in bullying their peers, such as the social isolation of the victim from his or her peer community, are largely inapplicable in the case of their teachers (see Dzuka & Dalbert, 2007). Further, according to Marian (2008), students reported having bullied their teachers by circulating insulting jokes and disrepectful public discussion of the private matters of teachers. In the study by James and colleagues (2008), some students reported having bullied their teachers by spreading rumors about them. In addition, the “sabotage” of teaching situations is a form of bullying quite often encountered by school teachers (De Wet, 2010a; James et al., 2008). Harassment by means of communication technology has also been mentioned as a form of bullying experienced by teachers (see Kauppi & Pörhölä, 2009). The studies concerning the bullying of teachers by students do not, however, present a similar spectrum of various forms of direct and indirect bullying as the studies focusing at bullying occurring between students (e.g., Pörhölä & Kinney, 2010; Rivers & Smith, 1994; Whitney & Smith, 1993). It is nevertheless probable that students bully their teachers in more versatile ways than what has been observed in earlier studies. In this study, we utilize a detailed measure focusing on the communicative features of bullying to survey the forms in which the bullying of teachers manifests itself.

PREVALENCE OF BULLYING AND CHARACTERISTICS OF STUDENTS WHO ENGAGE IN BULLYING BEHAVIOR Studies conducted in different countries (e.g., Dzuka & Dalbert, 2007; James et al., 2008; Khoury-Kassabri et al., 2009; Terry, 1998) have shown that bullying directed at teachers by

Teachers Bullied by Students

121

students is a significant concern for a large number of teachers at school. According to an extensive survey on the harassment and violence experienced by Finnish lower secondary school teachers by Salmi and Kivivuori (2009), 66% of the teachers surveyed had been subjected to insulting behavior by their students during their teaching career. Thirty percent had been subjected to harassment (for example, had had their property vandalized or their domestic privacy violated), 24% to the threat of violence, 11% to physical violence, and 8% to sexual harassment. Previous knowledge concerning the gender of the students who typically engage in bullying teachers is rather limited. Some studies have suggested that male students engage in more bullying and violent acts against teachers than female students (Chen & Astor, 2009; James et al., 2008; Khoury-Kassabri et al., 2009). In Finland, Kivivuori and Tuominen (1999) discovered that approximately three fourths of the students in primary and lower secondary schools who had subjected teachers to insulting communication were boys. Similarly, Rantala and Keskinen (2005) noted that 75% of lower secondary school teachers (not all of whom had personally experienced bullying) thought that it is particularly boys who act violently towards teachers. In addition, 79% of the respondents assessed that the students who had subjected teachers to harassment or violence had acted alone. Furthermore, Olweus (1993, 2003) noted that the students who act aggressively towards their fellow students often express aggression towards their teachers and parents as well. Olweus also assumed that these students would have a more positive than average attitude towards violent modes of behavior. Previously, most of the studies on students who bully teachers have been based on single incidents of violence or harassment. These studies have provided little information on the characteristics of the students who repeatedly utilize means of direct and indirect communication to bully their teachers. The existing studies do not shed any more light on the type of communication relationship that the teachers who experience bullying have with their bullies (see Kauppi & Pörhölä, 2009). For example, it is not known whether teachers are subjected to repeated bullying by the students with whom they work most often or by the students they know least well. Very little is also known about how the students who bully teachers behave in their own peer relationships. It is not known whether the same students who bully teachers also bully their fellow students or whether they are themselves bullied by their fellow students.

RESEARCH QUESTIONS The objective of this study is to examine the direct and indirect forms in which the bullying of teachers by students manifests itself, the characteristics (gender and number) of students who bully teachers, the relationship existing between teachers and the students who bully them, and how the students who bully teachers behave in their peer relationships. In this article, we utilize the following three research questions to examine the experiences of 70 teachers who had been subjected to bullying by their students:

1. What types of bullying are teachers exposed to by students? 2. In the experience of teachers, what type (what gender, which number) of students subject them to bullying, and what type of a relationship do teachers have with the students who bully them? 3. According to teachers’ perception, how do the students who bully teachers behave in their own peer relationships?

122

Kauppi and Pörhölä

METHODS Procedures The study was conducted as an Internet survey using SPSS mrInterview software. The data were gathered from school teachers in various parts of Finland during the school year 2008–2009. The survey was sent to 86 schools, among which were primary schools (students aged 7–13 years), lower secondary schools (students aged 13–16 years), and integrated comprehensive schools (students aged 7–16 years). The size of the schools ranged from three-teacher village schools to large comprehensive schools in cities of over 100,000 inhabitants. These schools employed a total of around 2,000 teachers, all of whom were informed of the survey by the principals of the schools. Before the questionnaire was sent out, school principals were asked to inform all the teachers in their school about the survey. In some cases, contact was first made with the director of education and cultural services of the city or municipality who then asked the principals to send the survey either to those schools that had agreed to participate in advance or to all schools in that city or municipality. The teachers were sent, by e-mail, a cover letter that included a link to the questionnaire. The letter also included information about the study and matters of confidentiality relating to the handling of the material. The teachers were asked to respond to the questionnaire within 3 weeks. Other instructions for the filling in of the questionnaire were provided in the web form. As the respondents were adults voluntarily participating in the study and the data was collected anonymously, according to the Finnish standards, approval of the study protocol by an ethics committee was not required. Instead, after the principal of each school had given his or her approval for the data collection in that particular school, voluntarily participating teachers gave their informed consent by responding to the questionnaire.

Respondents A total of 215 comprehensive school teachers responded to the survey. Of those respondents who identified their gender, 147 (76.2%) were female and 46 (23.8%) were male. Twenty-two of the respondents did not identify their gender. The gender distribution of the respondents corresponded well with the gender distribution of the teachers working in Finnish primary and lower secondary schools (with students aged 7–16 years). In Finland, approximately 74% of teachers at these school levels are female (Ojala, 2009). Fifty-five (25.6%) of the respondents reported occasionally having been subjected to bullying (as defined in the questionnaire form) by students, whereas seven teachers (3.3%) reported a bullying frequency of almost every week, and eight teachers (3.7%) reported having been bullied by students almost daily. One hundred and forty-five respondents (67.4%) reported hardly ever having been subjected to bullying by students. In this article, we look at teachers’ subjective experiences of being subjected to bullying. Therefore, the analysis of the material focuses on the respondents who had experienced bullying occasionally or more often (n 5 70). We will, hereafter, refer to them as either “respondents” or “teachers,” even though their number includes a few principals as well. The following description of the group of respondents is limited to these teachers only. Of those teachers who identified their gender, 52 (83.9%) were female and 10 (16.1%) were male. Eight of the respondents did not identify their gender. Eleven (15.7%) of the teachers were 20–30 years old, 19 (27.1%) were 31–40 years old, 29 (41.4%) were 41–50 years old, and 11 (15.7%) were 51–60 years old. Fourteen (20.0%) of

Teachers Bullied by Students

123

the ­respondents had worked as a teacher for 0–3 years, 20 (28.6%) for 4–10 years, 19 (27.1%) for 11–20 years, and 17 (24.3%) for more than 20 years. Among the teachers were 38 subject teachers (54.3%), 14 class teachers (20.0%), 11 special needs teachers (15.7%), 3 special class teachers (4.3%), 2 principals (2.9%), 1 part-time teacher (1.4%) and 1 teacher who reported a professional title of “class teacher and subject teacher” (1.4%). In Finnish primary and lower secondary schools, class teachers mainly work with students aged between 7 and 13 years, whereas subject teachers mainly work with students aged between 13 and 16 years.

Questionnaire and Analysis of Data For the purpose of the study, a questionnaire was developed that was used to assess teachers’ experiences of being subjected to bullying by students. We chose to use the term kiusaaminen (i.e., the Finnish equivalent to “bullying”) in the questionnaire because of its established use in Finnish studies concerning school and workplace bullying. The questionnaire consisted of several components. The bullying of teachers was defined as a communication process in which a teacher is repeatedly subjected, by one or more students, to interaction that he or she perceives as insulting, upsetting, or intimidating. Bullying can be verbal, nonverbal, or physical in nature. The teachers who reported having experienced bullying (as defined previously) either occasionally, almost every week, or almost daily were asked to answer to all questions in the questionnaire. The teachers who had not experienced bullying were asked to answer to five specific questions found in the questionnaire. Those questions are not discussed in this article. The following discussion of the questionnaire will only involve the components reported in this article. Teacher as a Victim of Bullying Scale. The survey of the manifestations of bullying experienced by teachers used a measure called “Teacher as a Victim of Bullying,” which was designed for this study. In this scale, the respondents who reported having experienced bullying were asked to think about a period of time during which they had been most intensely bullied by students and to describe the types of bullying they had been subjected to. To facilitate this, they were provided with a list covering 22 different forms of bullying (see Table 1). The measure was developed based on the existing knowledge of the forms in which the bullying of teachers manifests itself (see Kauppi & Pörhölä, 2009, for a review of literature). Where applicable, items of Pörhölä’s (2008) victimization scale designed to facilitate the surveying of bullying between students was also used in the development of the measure. Because previous studies on the bullying of teachers by students are scarce, additional components designed on the basis of studies on school and workplace bullying, as well as the rational reasoning of the researchers, were used in the development of the measure. The measure required the respondents to indicate the intensity of their experience of every form of bullying listed. The answer alternatives were hardly ever, occasionally, almost every week, and almost daily. The measure also featured an additional question that enabled the respondents to describe any forms of bullying not included in the list. When reporting the results, we examine the answers one component at a time, presenting frequencies and percentual frequencies. The reliability of the Teacher as a Victim of Bullying Scale was rather high (Cronbach’s alpha 5 .873; 22 items). Item–total correlations between the individual items and the scale varied from .12 to .66. Removal of only one item (i.e., harassment through e-mail,

124

Kauppi and Pörhölä

telephone calls, or text messages) had improved the reliability estimate a little. The theoretical components of the scale consisted of five categories of variables representing different types of bullying. These categories and the featured variables are described in the following section. A. Direct verbal bullying; 6 items (a 5 .738)   1. Denouncing or name-calling   2. Obscene or inappropriate comments   3. Making fun of, or laughing at the teacher (openly or behind his or her back)   4. Upsetting, sexually tinged remarks, or propositions   5. Unjust disparagement of the teacher’s professional skills  6. Threatening B. Direct nonverbal bullying; 3 items (a 5 .380)   7. Insulting gestures   8. Mimicking the teacher’s characteristic features (e.g., speech or walking style)   9. Violation of personal space or improper touching C. Physical bullying; 3 items (a 5 .721) 10. Physical violence 11. Violation of physical integrity committed with the intent of bullying (e.g., throwing of objects, pinching, or soiling of the teacher’s clothes) 12. Damaging or stealing of property D. Indirect private bullying; 6 items (a 5 .682) 13. Repeated lying to the teacher 14. Refusal to cooperate (e.g., repeated refusal to comply with the teacher’s instructions) 15.  Repeated disregard of the teacher’s presence (e.g., not responding to questions or requests) 16. Hiding from the teacher or repeatedly coming late to class 17. Harassment through e-mail, telephone calls, or text messages 18. Improper insinuation about the teacher’s private matters (e.g., matters relating to the teacher’s health or family) E. Indirect public bullying; 4 items (a 5 .702) 19. Making unfounded reports or complaints against the teacher 20. Spreading unfounded gossip 21. Bullying over the Internet (e.g., posting of abusive writings or images) 22. Subjecting the teacher to inappropriate attention in public places (e.g., through sassing, name-calling, wall writings)

The specification of different forms of bullying is based on the division between direct and indirect bullying generally made in research literature concerning school and workplace bullying. Direct bullying refers to the hurting of the victim through physical or verbal means, whereas in the case of indirect bullying, the victim is hurt in a less direct manner— for example, by presenting criticism behind his or her back or by spreading hurtful gossip and unfounded stories about him or her (e.g., Keashly & Jagatic, 2003; Rayner et al., 2002; Rivers & Smith, 1994; Whitney & Smith, 1993). We further divided the direct forms of bullying (variable categories A and B) into verbal and nonverbal varieties. We also looked at both direct and indirect physical bullying ­(variable category C). Furthermore, we divided the indirect forms of bullying into public

Teachers Bullied by Students

125

and private varieties on the basis of their external characteristics. By public indirect bullying (variable category E), we refer to communication that is intended to defame the teacher in the eyes of the school community or the wider public. Private indirect bullying (variable category D) differs from the public variety in that it does not involve actual verbal insults, insulting gesticulation, or attempts to defame the teacher; however, the students still communicate or act in ways that are perceived as insulting by the teacher. For the most part, the experience of being insulted originates from the teacher’s interpretation of a student’s behavior in a communication situation in which the student somehow breaks the communication norms or school rules. As far as we know, this model has not been employed in previous studies to categorize the different forms of bullying. Teachers’ experiences on being bullied by students were additionally assessed through one open-ended question. The respondents were asked to describe a typical situation where they had experienced being bullied by a student or students. Expressions describing the various forms of bullying were extracted from the answers and then assigned into categories utilizing a data-based classification method. The resulting categories were as follows: verbal insulting (e.g., sassing, name-calling, laughing at the teacher), breaking of communication norms (e.g., failure to greet the teacher, disregarding the teacher’s presence, mimicking), lying, violation of the teacher’s physical integrity, disparagement of the teacher’s professional skills, resisting the teacher’s orders (e.g., refusing to leave the class or to comply with regulations concerning the maintenance of order), repeatedly coming late to class, threatening, public abuse (e.g., posting writings on the Internet, writing things on walls), and not letting the teacher work in peace. When reporting the results, we examine these responses, especially insofar as they complement the responses provided for the Teacher as a Victim of Bullying Scale. Assessment of the Characteristics of Students Who Engage in Bullying and of the Relationship Between Teachers and Those Students. The following two structured questions were used to probe the teachers’ experiences of the students who bully them: “How many students typically take part in bullying you?” with the answer alternatives being “1”, “2 to 5”, “6 to 10” and “11 or more”; and “What gender are the students who bully you?”, to which the respondents could also answer by selecting the option “Both girls and boys”. The relationship between the teachers and the students who bully them was probed with the following question: “Are your bullies students whom you currently teach or have taught in the past?” The answer alternatives were as follows: (a) “students whom you currently teach,” (b) “students whom you have taught in the past,” and (c) “students whom you have never taught.” The respondents were given an opportunity to select more than one alternative. In the following section, we analyze the responses to these questions through frequencies and percentual frequencies. Teachers’ Conception of How the Students Who Bully Them Behave in Their Own Peer Relationships. The teachers were also asked to assess how the students who bully them behave in their own peer relationships. The question asked the respondents to indicate what types of students had bullied them. The following answer alternatives were provided: (a) “students who also regularly bully other students of the school,” (b) “students who also occasionally bully other students of the school,” (c) “students who are regularly bullied at school,” (d) “students who are occasionally bullied at school,” (e) “students who are bullied at school and who also bully other students,” (f) “students who do not bully other students and who are not bullied by others,” and (g) “other type, which?” In the next section, we analyze the responses to these questions through frequencies and percentual frequencies.

126

Kauppi and Pörhölä

RESULTS Firstly, we listed the types of bullying that the teachers had been subjected to by the students. Table 1 presents the frequencies by response category. We also formed a combined category that displays, in order of frequency, the frequencies and percentual frequencies for respondents who had experienced each form of bullying at least occasionally. As the last column of Table 1 reveals, the most common form of bullying encountered by the teachers was the utterance of obscene or inappropriate comments. The second and third most common forms of bullying were refusal to cooperate and repeated lying to the teacher, respectively. The table furthermore shows that the teachers reported having quite frequently been subjected to the following forms of bullying by students: making fun of or laughing at the teacher, denouncing and name-calling, insulting gesticulation, hiding from the teacher or coming late to class, unjust disparagement of the teacher’s professional skills, and mimicking of the teacher’s communication. When the results are examined from the point of view of the types of bullying that teachers had experienced at least once a week, the most common forms of bullying were refusal to cooperate (once a week or more frequently, f 5 25; 36.2%) and hiding from the teacher or repeatedly coming late to class (once a week or more frequently, f 5 19; 27.5%). Hiding from the teacher and repeatedly coming late to class are examples of indirect communication that, if recurrent, may be interpreted as bullying by the teacher. The results showed that all of the components of the scale described a form of bullying that the respondents had at least occasionally experienced at their work. The results also indicate that teachers are typically subjected to bullying during the school day. Technologically mediated bullying appeared to be relatively rare but nevertheless occurred. The respondents were also given an opportunity to report any forms of bullying they had experienced that were not listed in the components of the scale. In these responses, the teachers most frequently described a situation where a student’s parents had behaved inappropriately towards them. Forms of bullying mentioned in individual responses also included running away from the teacher (with the intent to irritate him or her) and engaging in disorderly behavior in class. A female teacher had experienced as bullying the wish, expressed by a student, that she would be replaced by a male teacher. The respondents were also asked to describe, in their own words, a typical situation where they had experienced being bullied by students. A total of 61 teachers answered to the question. The majority of them described either a situation where a student or students had insulted them verbally or laughed at them (20 teachers) or a situation where they had not been allowed to work in peace in the classroom (20 teachers). The following response describes an example of the latter (citations are translated freely from Finnish into English by authors): “There are a few students who are intent on transforming the learning situation in the classroom into a chaotic, uncontrollable situation, where the focus would be on having fun and talking (hollering) to one’s friends” (subject teacher, 20–30 years old). The potential subtlety of bullying, and the difficulty of its interpretation is not only interesting but also quite illustrative of the forms of bullying experienced by teachers in their work. One of the respondents stated that the experience of being bullied can be almost impossible to explain: “Sometimes the bullying can be so subtle that I would be hard pressed to explain what the student actually does, but it still really gets my goat” (subject teacher, 20–30 years old). The forms of bullying described by other respondents were similar to those described in the components of the measure.

Teachers Bullied by Students

127

TABLE 1.  The Forms of Bullying Experienced by the Teachers (%5 )

The Forms of Bullying

n

f1

f2

f3

f4

f5

Obscene or inappropriate comments

70

 7

44

10

 9

63 (90.0)

Refusal to cooperate

69

10

34

12

13

59 (85.5)

Repeated lying

69

13

39

10

 7

56 (81.2)

Making fun of or laughing at the teacher

69

14

39

10

 6

55 (79.7)

Denouncing or name-calling

69

21

37

4

 7

48 (69.6)

Insulting gestures

69

22

32

11

 4

47 (68.1)

Hiding or repeatedly coming late to class

69

26

24

14

 5

43 (62.3)

Unjust disparagement of the teacher’s   professional skills

69

26

34

 6

 3

43 (62.3)

Mimicking of the teacher’s characteristic  features

69

29

35

 4

 1

40 (58.0)

Repeated disregard of the teacher’s presence

69

33

27

 4

 5

36 (52.2)

Threatening

68

37

26

 2

 3

31 (45.6)

Violation of physical integrity (e.g., throwing   of objects or soiling the clothes)

68

47

16

 3

 2

21 (30.8)

Damaging or stealing of property

68

47

21

—

—

21 (30.8)

Subjecting the teacher to inappropriate   attention in public places

68

48

16

 3

 1

20 (29.4)

Spreading unfounded gossip

69

49

17

 2

 1

20 (29.0)

Improper insinuation about the teacher’s   private matters

68

49

18

 1

—

19 (27.9)

Physical violence

68

50

13

 3

 2

18 (26.5)

Making unfounded reports or complaints   against the teacher

66

50

16

—

—

16 (24.2)

Irritating sexually tinged remarks or  propositions

68

56

11

 1

—

12 (17.6)

Violation of personal space or improper  touching

67

57

 7

 3

—

10 (14.9)

Harassment through e-mail, telephone calls or   text messages

68

58

10

—

—

10 (14.7)

Bullying over the Internet (e.g., posting of   abusive writings or images)

68

63

 4

 1

—

 5  (7.6)

Note. 1Hardly ever; 2Occasionally; 3Almost weekly; 4Almost daily; 5Combined occasionally or more often.

128

Kauppi and Pörhölä

TABLE 2.  The Gender of the Students Who Had Bullied Teachers The Gender of Students

f (%)

Female

  4 (5.7)

Male

44 (62.9)

Both female and male

22 (31.4)

n

70 (100.0)

Secondly, we focused on determining the gender distribution of the students who had bullied teachers. As indicated in Table 2, the majority (63%) of the teachers reported that their bullies had been male students. Of the respondents, 31% reported having been bullied by both male and female students, and only 6% stated that they had been bullied exclusively by female students. Thirdly, we examined the teachers’ statements concerning the number of the students who took part in the bullying. As Table 3 indicates, this number was usually quite small. Of the teachers, 46% reported that they were typically bullied by only one student. Another 46% of the respondents stated that the number of students who took part in the bullying ranged between two and five. Only 9% of the respondents reported that more than five students had taken part in the bullying. Fourthly, we examined the relationship between the teachers and the students who bullied them. The results showed that the teachers had most frequently experienced bullying by students whom they were teaching at the time of their response. This answer alternative was selected by 66% of the teachers. Of the respondents, 36% stated that they had been bullied by students whom they had taught in the past. Notably, as many as 24% of the teachers reported having been subjected to bullying also or exclusively by students whom they had never taught. The respondents were given an opportunity to select more than one alternative. Fifthly, the teachers were asked to assess how the students who had bullied them behave in their own peer relationships. The respondents were given an opportunity to select more than one alternative (see Table 4). As Table 4 shows, a significant majority of the teachers assessed that the students who had bullied them also at least occasionally bullied their fellow students. Furthermore, one tenth of the respondents thought that the students who had bullied them were themselves TABLE 3.  The Number of the Students Who Took Part in Bullying Teachers The Number of Students

f (%)

1

32 (45.7)

2–5

32 (45.7)

6–10

  4 (5.7)

more than 10

  2 (2.9)

n

70 (100.0)

Teachers Bullied by Students

129

TABLE 4.  The Teachers’ Conception of How the Students Who Had Bullied Them Behave in Their Own Peer Relationships Selected Alternatives f

%a

They also regularly bully other students of the school.

 35

50.0

They also occasionally bully other students of the school.

 21

30.0

They are regularly subjected to bullying at school.

  2

 2.9

They are occasionally subjected to bullying at school.

  7

10.0

They are subjected to bullying and also bully other students.

  7

10.0

They do not bully other students, nor are they subjected to  bullying.

  3

 4.3

Other type of behavior, which?

 12

17.1

Cannot say.

 14

20.0

Total number of alternatives selected

101

Students’ Behavior in Their Peer Relationships

aPercentage

of respondents who selected the alternative (n 5 70).

subjected to bullying at school but also engaged in bullying their fellow students. Only a few teachers believed that they were typically bullied by students who neither bullied their fellow students nor were subjected to bullying themselves. The answers to the question “Other type, which?” indicated that the respondents believed that the students who had bullied them bullied other teachers of the school as well.

DISCUSSION The focus of this study was to describe the special characteristics of the bullying of ­teachers by students. The teachers who responded to the survey had been subjected to both direct and indirect forms of verbal and nonverbal bullying by students in their work. Direct forms of bullying typically included obscene and inappropriate commentary, making fun of and laughing at the teacher, denouncing, name-calling, and insulting gesticulation. Typical indirect forms of bullying included refusal to cooperate with the teacher and not allowing the teacher to work in peace during teaching situations. These forms of bullying have also been reported in earlier studies (e.g., De Wet, 2010a; Dzuka & Dalbert, 2007; James et al., 2008; Kivivuori & Tuominen, 1999; Terry, 1998). The respondents’ bullying experiences additionally covered a wide array of various forms of insulting behavior not identified in earlier studies (cf. De Wet, 2010a; James et al., 2008; Kauppi & Pörhölä, 2009; Terry, 1998). These include repeated lying, ­mimicking the teacher’s speech or walking style, and repeatedly coming late to class and hiding from the teacher. This type of communication represents a type of violation of communication norms one would not necessarily expect to encounter in the role of a teacher. Disturbing class, repeatedly coming late to class, and hiding from the teacher are forms of bullying

130

Kauppi and Pörhölä

that can be used to prevent teachers from performing their duties to their own satisfaction and from fulfilling the requirements of their position. Although, lateness and absenteeism by students, for example, have been recognized to be a problem for teachers in surveys concerning students’ disruptive behavior (e.g., Teachers’ Union of Ireland [TUI], 2006), these kind of behaviors have not been examined as teachers’ bullying experiences in previous research. All in all, it would appear that the direct forms of bullying experienced by teachers are notably similar to the direct forms of bullying to which students subject their fellow students (e.g., Olweus, 1993; Pörhölä & Kinney, 2010; Rivers & Smith, 1994). However, the indirect forms of bullying experienced by teachers somewhat differ from the forms of indirect bullying to which students subject their fellow students (e.g., exclusion of a fellow student from group; see Olweus, 1993; Rivers & Smith, 1994). The indirect forms of bullying experienced by teachers and also some of the direct forms of bullying such as unjust disparagement of teachers’ professional skills rather contain typical characteristics of workplace bullying (see, e.g., Keashly & Jagatic, 2003; Pörhölä et al., 2006, for literature reviews), although the nature of teacher’s work brings certain special characteristics to the bullying. It appears that teachers are also subjected to physical bullying as well as technologically mediated bullying, although these occur clearly less frequent than other forms of bullying. The teachers who had experienced bullying stated that they had been mostly bullied by male students. This observation supports the findings from previous studies (e.g., James et al., 2008; Khoury-Kassabri et al., 2009; Kivivuori & Tuominen, 1999; Rantala & Keskinen, 2005). Although research has already shown that when a student bullies fellow students, the bully is more likely to be male than female (e.g., Fekkes, Pijpers, & VerlooveVanhorick, 2005; Luopa, Pietikäinen, & Jokela, 2008; Scheithauer, Hayer, Petermann, & Jugert, 2006); the role of boys appears to become more pronounced when the object of the bullying is a teacher. As Dzuka and Dalbert (2007) also noted, it is probable that indirect bullying, more typically employed by girls (see Pörhölä & Kinney, 2010), is more difficult to direct at a teacher. Further studies should then focus on examining whether the bullying of teachers by girls differs in nature from the bullying of teachers by boys. The teachers had most frequently experienced bullying by just one student or a small group of students. Only a few had been bullied by a group of more than five students. This observation differs from those made on bullying occurring between students. Research has shown that when bullying occurs between students, a large number of students take part in the bullying in various roles (e.g., Salmivalli, Lagerspetz, Björkqvist, Österman, & Kaukiainen, 1996). Although the bullying of teachers by students does not appear to contain the characteristics of a group phenomenon so evident as in the bullying of students by students, future studies should examine in more details, the number of bullies and the possible participation roles of students in bullying situations where the object of bullying is a teacher. It was discovered that the teachers were most frequently subjected to bullying by students whom they were teaching at the time of their response. This result is easy to understand because teachers spend most of their working time with their own students. Considering this, it was interesting to observe that as many as one fourth of the respondents had experienced bullying by students whom they had never taught. Particularly in the case of large schools, it is probable that the teacher and the students who bully him or her do not always really know each other and, thus, do not always share an actual ­communication relationship. It can be asked, then, what purpose does it serve for the bullies to bully a teacher who is nearly unknown to them. It is clear that in such a case, bullying cannot be

Teachers Bullied by Students

131

related to any problems existing in the communication relationship between the parties. It is possible that by bullying an unknown teacher, the bullies aim to gain more power and improve their status among other students. The teacher’s role in this process is just that of a tool used to achieve this end. A significant majority of the teachers assessed that the students who had bullied them also bullied their fellow students at school. This means that bullying was seen as a typical model of behavior for those students. Future studies should set out to examine whether some students are more inclined to bully both their fellow students and teachers and whether the motivation behind the bullying remains unchanged for such students regardless of the position of the individual they choose to bully. It should also be examined whether the means, which such students employ in bullying others, vary according to the object. This type of understanding would shed light into whether the selection of the forms of bullying actually employed is more dependent on the characteristics and communicative traits of the bully or on those of the victim. A good understanding of the special characteristics of the phenomenon of bullying of teachers by students is particularly needed in teacher training.

LIMITATIONS The selection of schools for this study was not based on systematic sampling. Instead, the request to participate and the questionnaire were sent to teachers working in a number of preselected schools that had expressed their willingness to participate in the study. It is probable that the study elicited most responses from teachers who had experienced bullying by students. This may have been one of the reasons for the low response percentage. Be that as it may, no conclusions about the prevalence of bullying of teachers by students can be drawn based on the material nor was the goal of this study. The main goal of the study was to explore the nature of bullying experienced by teachers and the characteristics of the students who bullied them. Even though we did not use systematic sampling to collect a representative sample of teachers working in Finnish schools, we believe that responses of the 70 teachers who indicated that they had been bullied by their students enabled reaching these goals. Despite the fact that this was not a random sample, the gender distribution of the 215 respondents corresponded well with the gender distribution of the teachers working in Finnish primary and lower secondary schools (see Ojala, 2009). Similarly, the age distribution of the respondents corresponded quite well with the age distribution of Finnish teachers: 35.3% of teachers working in Finnish primary and lower secondary schools are aged younger than 40 years; in the data at hand, the proportion of teachers aged between 20 and 40 years was 39.3%. Our material was slightly biased towards respondents aged between 41 and 50 years; their proportion in our material was approximately 10 percentage points larger than the proportion of that age group among primary and lower secondary school teachers. Correspondingly, the proportion of respondents aged older than 50 years in our material was slightly over 10 percentage points smaller than the proportion of that age group among teachers working in Finnish primary and lower secondary schools. Schools of different sizes and located in cities as well as the countryside were represented in the study. However, because of the low response percentage, the results cannot be generalized to all teachers in Finland. In future studies, using representative cultural or cross-cultural samples, more attention could be paid to individual and school-level variability in teachers’ experiences of bullying.

132

Kauppi and Pörhölä

REFERENCES Basic Education Act (2003/477, 7:36§, 7:36a§, 7:36b§). Retrieved March 30, 2011, from http:// www. finlex.fi/fi/laki/ajantasa/1998/19980628 Benefield, J. (2004). Teachers—the new targets of schoolyard bullies? Paper presented to New Zealand Association for Research in Education (NZARE). Retrieved March 8, 2011, from http://ppta.org.nz/index.php/collectiveagreements/hours-of-work/doc_details/156-teachers-thenew-targets-of-schoolyard-bullies Björkqvist, K., Österman, K., & Hjelt-Bäck, M. (1994). Aggression among university employees. Aggressive Behavior, 20, 173–184. http://dx.doi.org/10.1002/1098-2337(1994)20:33.0.CO;2-D Blase, J. J., & Blase, J. R. (2003). Breaking the silence: Overcoming the problem of principal mistreatment of teachers. Thousand Oaks, CA: Corwin Press. Cemaloglu, N. (2007). The exposure of primary school teachers to bullying: An analysis of various variables. Social Behavior and Personality, 35, 789–802. http://dx.doi.org/10.2224 /sbp.2007.35.6.789 Chan, J. H. F. (2009). Where is the imbalance? Journal of School Violence, 8(2), 177–190. http:// dx.doi .org/10.1080/15388220802074199 Chen, J. K., & Astor, R. A. (2009). Students’ reports of violence against teachers in Taiwanese schools. Journal of School Violence, 8(1), 2–17. http://dx.doi.org/10.1080/15388220802067680 De Wet, C. (2010a). Victims of educator-targeted bullying: A qualitative study. South African Journal of Education, 30, 189–201. De Wet, C. (2010b). The reasons for and the impact of principal-on-teacher bullying on the victims’ private and professional lives. Teaching and Teacher Education, 26, 1450–1459. http://dx.doi .org/10.1016/j.tate.2010.05.005 Dworkin, A. G., Haney, C. A., & Telschow, R. L. (1988). Fear, victimization, and stress among urban public school teachers. Journal of Organizational Behavior, 9, 159–171. Dzuka, J., & Dalbert, C. (2007). Student violence against teachers: Teachers’ well-being and the belief in a just world. European Psychologist, 12, 253–260. http://dx.doi.org/10.1027/10169040.12.4.253 Einarsen, S., Hoel, H., Zapf, D., & Cooper, C. L. (2003). The concept of bullying at work: The European tradition. In S. Einarsen, H. Hoel, D. Zapf, & C. L. Cooper (Eds.), Bullying and emotional abuse in the workplace: International perspectives in research and practice (pp. 3–30). London: Taylor & Francis. Fekkes, M., Pijpers, F. I. M., & Verloove-Vanhorick, S. P. (2005). Bullying: Who does what, when and where? Involvement of children, teachers and parents in bullying behavior. Health Education Research, 20(1), 81–91. http://dx.doi.org/10.1093/her/cyg100 Fisher, K., & Kettl, P. (2003). Teachers’ perceptions of school violence. Journal of Pediatric Health Care, 17(2), 79–83. http://dx.doi.org/10.1067/mph.2003.20 Hoel, H., Faragher, B., & Cooper, C. L. (2004). Bullying is detrimental to health, but all bullying behaviours are not necessarily equally damaging. British Journal of Guidance & Counselling, 32, 367–387. http://dx.doi.org/10.1080/03069880410001723594 James, D. J., Lawlor, M., Courtney, P., Flynn, A., Henry, B., & Murphy, N. (2008). Bullying behaviour in secondary schools: What roles do teachers play? Child Abuse Review, 17, 160–173. http://dx.doi.org/10.1002/car.1025 Kauppi, T., & Pörhölä, M. (2009). Harassment experienced by school teachers from students: A review of the literature. In T. A. Kinney & M. Pörhölä (Eds.), Anti and pro-social ­communication: Theories, methods and applications (pp. 49–58). New York: Peter Lang. Keashly, L., & Jagatic, K. (2003). By any other name: American perspectives on workplace bullying. In S. Einarsen, H. Hoel, D. Zapf, & C. L. Cooper (Eds.), Bullying and emotional abuse in the workplace: International perspectives in research and practice (pp. 31–61). London: Taylor & Francis.

Teachers Bullied by Students

133

Keashly, L., & Nowell, B. L. (2003). Conflict, conflict resolution and bullying. In S. Einarsen, H. Hoel, D. Zapf, & C. L. Cooper (Eds.), Bullying and emotional abuse in the workplace: International perspectives in research and practice (pp. 339–358). London: Taylor & Francis. Khoury-Kassabri, M., Astor, R. A., & Benbenishty, R. (2009). Middle Eastern adolescents’ perpetration of school violence against peers and teachers: A cross-cultural and ecological analysis. Journal of Interpersonal Violence, 24(1), 159–182. http://dx.doi.org/10.1177/0886260508315777 Kivivuori, J., & Tuominen, M. (1999). Oppilaiden opettajiin kohdistama häirintä ja väkivalta Helsingin suomenkielisissä kouluissa [Harassment and violence directed at teachers by ­students in Finnish-language schools in Helsinki]. In J. Kivivuori, M. Tuominen, & K. Aromaa, “Mä teen mitä mä haluun”: Oppilaiden opettajiin kohdistama häirintä ja väkivalta Helsingin ­kouluissa 1997–1998 (Research Communications No. 44, pp. 9–55). Helsinki, Finland: National Research Institute of Legal Policy. Ojala, M. L. (2009). Perusopetuksen ja lukiokoulutusen opettajat kevätlukukaudella 2008 [Teachers in primary, lower secondary, and high schools in Spring 2008]. In T. Kumpulainen (Ed.), Opettajat Suomessa 2008 (pp. 29–56). Helsinki, Finland: Opetushallitus utbildningsstyrelsen. Retrieved March 30, 2011, from http://www.oph.fi/julkaisut/2009/Opettajat_Suomessa_2008.pdf Lahelma, E., Palmu, T., & Gordon, T. (2000). Intersecting power relations in teachers’ experiences of being sexualized or harassed by students. Sexualities, 3, 463–481. http://dx.doi.org/ 10.1177/136346000003004006 Leymann, H. (1996). The content and the development of mobbing at work. European Journal of Work and Organizational Psychology, 5, 165–184. http://dx.doi.org/10.1080/13594329608414853 Leymann, H., & Gustafsson, A. (1996). Mobbing at work and the development of post-traumatic stress disorders. European Journal of Work and Organizational Psychology, 5, 251–276. http://dx .doi.org/10.1080/13594329608414858 Luopa, P., Pietikäinen, M., & Jokela, J. (2008). Koulukiusaaminen peruskoulun yläluokilla 2000– 2007 [School bullying at lower secondary schools 2000–2007] (Opetusministeriön julkaisuja 2008:7). Retrieved September 12, 2009, from http://www.minedu.fi/export/sites/default/OPM/ Julkaisut/2008/liitteet/opm07.pdf?lang=fi Marian, C. (2008, July). Pattern of aggressive behaviour between teacher and student in school context. Paper presented at the world meeting of International Society for Research on Aggression, Budapest, Hungary. Matthiesen, S. B., & Einarsen, S. (2004). Psychiatric distress and symptoms of PTSD among victims of bullying at work. British Journal of Guidance & Counselling, 32, 335–356. http://dx.doi .org/10.1080/03069880410001723558 Olweus, D. (1993). Bullying at school: What we know and what we can do. Oxford, United Kingdom: Blackwell. Olweus, D. (2003). Bully/victim problems in school: Basic facts and an effective intervention ­programme. In S. Einarsen, H. Hoel, D. Zapf, & C. L. Cooper (Eds.), Bullying and emotional abuse in the workplace: International perspectives in research and practice (pp. 62–78). London: Taylor & Francis. Pörhölä, M. (2008). Koulukiusaaminen nuoren hyvinvointia uhkaavana tekijänä. Miten käy kiusatun ja kiusaajan vertaissuhteille? [School bullying as a threat to young people’s well-being: What happens to the peer relationships of the bullied and the bullies?] In M. Autio, K. Eräranta, & S. Myllyniemi (Eds.), Polarisoituva nuoruus? Nuorten elinolot-vuosikirja 2008 (pp. 94–104). Helsinki, Finland: Nuorisotutkimusverkosto/ Nuorisotutkimusseura, Nuorisoasiainneuvottelukunta, & Stakes. Pörhölä, M., Karhunen, S., & Rainivaara, S. (2006). Bullying at school and in the workplace: A challenge for communication research. In C. S. Beck (Ed.), Communication yearbook 30 (pp. 249–301). Mahwah, NJ: Erlbaum. Pörhölä, M., & Kinney, T. A. (2010). Bullying: Contexts, consequences, and control. Barcelona, Spain: Editorial Aresta. Rantala, T., & Keskinen, S. (2005). “Ei meidän koulussa . . . eihän?” Opettajat oppilaiden k­ iusaamina [Not in our school . . . is it? Teachers as victims of harassment by students]. In S. Keskinen

134

Kauppi and Pörhölä

(Ed.), Valta, kilpailu ja kiusaaminen opettajan työssä: Artikkelisarja (pp. 120–160). Helsinki, Finland: Opetus-, kasvatus- ja koulutusalojen säätiö OKKA. Rayner, C., Hoel, H., & Cooper, C. L. (2002). Workplace bullying: What we know, who is to blame, and what can we do? New York: Taylor & Francis. Rayner, C., & Keashly, L. (2005). Bullying at work: A perspective from Britain and North America. In S. Fox & P. E. Spector (Eds.), Counterproductive work behavior: Investigations of actors and targets (pp. 271–296). Washington, DC: American Psychological Association. Rivers, I., & Smith, P. K. (1994). Types of bullying behaviour and their correlates. Aggressive Behavior, 20, 359–368. http://dx.doi.org/10.1002/1098-2337(1994)20:53.0 .CO;2-J Salmi, V., & Kivivuori, J. (2009). Opettajiin kohdistuva häirintä ja väkivalta 2008 (Verkkokatsauksia 10/2009) [Harassment and violence directed at teachers 2008]. Helsinki, Finland: National Research Institute of Legal Policy. Retrieved June 29, 2009, from http://www.optula.om.fi/ uploads/u1sba5z5n9.pdf Salmivalli, C., Lagerspetz, K., Björkqvist, K., Österman, K., & Kaukiainen, A. (1996). Bullying as a group process: Participant roles and their relations to social status within the group. Aggressive Behavior, 22, 1–15. http://dx.doi.org/10.1002/(SICI)1098-2337(1996)22:13.0.CO;2-T Scheithauer, H., Hayer, T., Petermann, F., & Jugert, G. (2006). Physical, verbal, and relational forms of bullying among German students: Age trends, gender differences, and correlates. Aggressive Behavior, 32, 261–275. http://dx.doi.org/10.1002/ab.20128 Smith, P. K. (2012). Cyberbullying and cyber aggression. In S. R. Jimerson, A. B. Nickerson, M. J. Mayer, & M. J. Furlong (Eds.), Handbook of school violence and school safety: International research and practice (2nd ed., pp. 93–104). New York: Routledge. Teachers’ Union of Ireland. (2006). Findings of TUI survey on second level classroom disruption. Retrieved March 30, 2011, from http://www.tui.ie/_fileupload/Image/DisciplineSurvey.doc Terry, A. A. (1998). Teachers as targets of bullying by their pupils: A study to investigate incidence. British Journal of Educational Psychology, 68, 255–268. Twemlow, S. W., Fonagy, P., Sacco, F. C., & Brethour, J. R., Jr. (2006). Teachers who bully students: A hidden trauma. International Journal of Social Psychiatry, 52(3), 187–198. http://dx.doi .org/10.1177 /0020764006067234 Van Dick, R., & Wagner, U. (2001). Stress and strain in teaching: A structural equation approach. British Journal of Educational Psychology, 71, 243–259. http://dx.doi.org/10.1348/ 000709901158505 Whitney, I., & Smith, P. K. (1993). A survey of the nature and extent of bullying in junior/ middle and secondary schools. Educational Research, 35, 3–25. http://dx.doi.org/10.1080/ 0013188930350101 Acknowledgments. The article is based on the doctoral dissertation of the first author Teemu Kauppi, directed by the second author Maili Pörhölä. Preparation of this article was supported by Finnish Cultural Foundation and Finnish Work Environment Fund (Grant No. 107341) in the case of the first author, the Academy of Finland (project no. 106221) in the case of the second author, and University of Jyväskylä, Finland, for both authors. The main results have previously been published in Finnish language in a periodical (Työelämän tutkimus, 2010, 8) with limited circulation in Finland. Correspondence regarding this article should be directed to Teemu Kauppi, MA, or Maili Pörhölä, PhD, Department of Communication, University of Jyväskylä, Finland, P.O. Box 35, FI-40014. E-mail: [email protected] or [email protected]

Violence and Victims, Volume 22, Number 6, 2007

Perpetrators and Targets of Bullying at Work: Role Stress and Individual Differences Stig Berge Matthiesen, PhD Ståle Einarsen, PhD University of Bergen, Norway A workplace survey study (N = 2215, response rate 47%) revealed that about 16% of the sample may be categorized as either perpetrators (5.4%), provocative victims (2.1%), or as targets of bullying (8.3%). Targets of bullying, provocative victims, and bullies were compared with those 84% who do not report any involvement with respect to bullying at work, self-esteem, aggressive tendencies, prior experiences of bullying, or experiences of role stress. Perpetrators were found to have a higher level of aggression than did the comparison group and the targets. Provocative victims manifested a low level of self-esteem and social competency combined with a high level of aggressiveness. Targets of bullying revealed low levels of self-esteem and social competency. Targets, provocative victims, and perpetrators reported elevated levels of role stress in the form of unclear or conflicting demands and expectations around work tasks and daily work.

Keywords: bullying at work; harassment at work; provocative victims; bullies; personality; workplace aggression

B

ullying is considered to be a subset of the overarching concept of aggression (Griffin & Gross, 2004). The first empirical studies in which the term “bullying” or “mobbing” was applied without referring to schoolyard bullying were published in Scandinavia around 1990 (Leymann, 1990; Matthiesen, Raknes, & Rokkum, 1989). Since then, several studies have shown that bullying is a widespread phenomenon in many countries, and large-scale studies in Scandinavia have indicated that approximately 3%–4% of the working population are exposed to this kind of misbehavior (Einarsen & Skogstad, 1996; Leymann, 1992). Finnish and British studies have revealed even higher prevalence rates of approximately 10% (Hoel & Cooper, 2000; Vartia, 1996). Since the onset of this research, the concept of bullying at work has been considered a synonym for the concepts mobbing and harassment at work (Einarsen, Hoel, Zapf, & Cooper, 2003). The Scandinavian approach defines bullying as a situation in which one or more persons systematically and over a long period of time perceive themselves to be on the receiving end of negative treatment on the part of one or more persons, in a situation in which the person(s) exposed to the treatment have difficulty in defending themselves against this treatment (Einarsen, 2000). This definition is adapted from research on bullying among schoolchildren (Olweus, 1978, 1993, 2003). Keashly and Jagatic (2003) defined bullying as interactions between organizational members that are characterized by repeated hostile © 2007 Springer Publishing Company Company

135 735

736 136

Matthiesen and Einarsen

verbal and nonverbal behavior, often nonphysical, directed at a person in such a way that the target’s sense of himself as a competent worker and person is negatively affected. “Bullying can be considered as a form of coercive interpersonal influence. It involves deliberately inflicting injury or discomfort on another person repeatedly through physical contact, verbal abuse, exclusion, or other negative actions” (Forsyth, 2006, p. 206). Hence, bullying is a long-lasting phenomenon that “wears down” its victims (Einarsen & Skogstad, 1996). Typically the bullying process lasts for more than a year. Previous studies on bullying have revealed approximately equal victimization rates among men and women (Einarsen & Skogstad, 1996; Niedl, 1995; Vartia, 1996). Others have estimated that about two-thirds of the targets are women, while 50%–80% of the perpetrators are managers, most often men (Zapf, Einarsen, Hoel, & Vartia, 2003). Many of the targets of bullying suffer from severe health problems, such as depression, anxiety, compulsive behavior (Matthiesen & Einarsen, 2001; Mikkelsen & Einarsen, 2002b; Niedl, 1996), or posttraumatic stress symptoms (Leymann & Gustavson, 1996; Matthiesen & Einarsen, 2004; Mikkelsen & Einarsen, 2002a; Niedl, 1996; Nielsen, Matthiesen, & Einarsen, 2005). So far there has been comparatively more research on the consequences of bullying than on its antecedents (Einarsen et al., 2003). An especially controversial issue in the field has been the role of personality characteristic as antecedents of bullying behaviors and experiences of victimization from bullying (Zapf & Einarsen, 2003).

PERPETRATORS OF BULLYING Important empirical knowledge about the phenomenon of workplace bullying has been collected throughout the last decade (Einarsen et al., 2003). However, a paradox exists, which Rayner and Cooper (2003) refer to as a “black hole” in the research field. Perpetrator behavior and perpetrator characteristics have generally been reported by the targets of bullying (e.g., Adams, 1992; Kile, 1990), or in anecdotal stories in more popular books (e.g., Bing, 1992). There are only a few published studies in which the perpetrators of bullying have been the subjects of empirical research. Some exceptions do exist, however. A survey study conducted by Coyne Chong, Seigne, and Randall (2003) revealed that 19.3% of a 288 personnel sample (members of various work groups) indicated that they had subjected others to bullying. This somewhat surprisingly high percentage of bully behavior decreased to 2.7%, however, when the role as a perpetrator was defined more strictly operationally, namely as a combination of self-report and peer-report (i.e., the person himself admits bullying of others, a confession validated by at least two colleagues). The self-reported and peer-reported group of perpetrators was found to be different from the control sample in terms of the personality factor mental stability. When personality dimensions such as independence, conscientiousness, and extraversion were assessed, minor and nonsignificant group differences were found. According to victim reports, perpetrators are male more often than female, and supervisors and managers more often than colleagues (Zapf & Einarsen, 2003; Zapf et al., 2003). Summarizing the sparse empirical findings on the bully as a cause of workplace bullying, Zapf and Einarsen (2003) have suggested three main antecedents of bullying related to perpetrator characteristics: (a) self-regulatory processes with regard to threatened self-esteem, (b) lack of social competencies, and (c) bullying as a result of what has been labeled as micropolitical behavior (Neuberger, 1989), that is, internal rivalry or competition in the workplace, particularly when there is a lack of formal structures and clearly divided work tasks and responsibilities.

Bullying at Work 

737 137

Self-Esteem and Social Competence In a review of self-esteem research, Baumeister, Smart, and Boden (1996) proposed that it is high self-esteem that is related to aggressive behavior. Low self-esteem is linked to depressive reactions and withdrawal. Individuals with low self-esteem are therefore rarely aggressive because they fear losing the encounter (Zapf & Einarsen, 2003). Individuals with low self-esteem experience self-doubt, anxiety, self-contempt, and ultimately depression (Beck, Rush, Shaw, & Emery, 1979), that is, various reactions that on the individual level are inner (intrinsic) directed, against the persons themselves. A high level of self-esteem is not entirely positive, however. High self-esteem can constitute a stable or an unstable selfevaluation. People with unstable high self-esteem may well become aggressive in response to even seemingly minor or trivial threats to self-esteem (Zapf & Einarsen, 2003), for example, after receiving unfavorable feedback (Kernis, Cornell, Sun, Berry, & Harlow, 1993). Bullying may be regarded as a more external (extrinsic) reaction directed against some part of the daily surroundings. Thus, a high level of self-esteem can lead to external reactions such as facing others with tyrannical behavior. Correspondingly, high self-esteem has been found to be related to perfectionism, arrogance, and narcissism (Ashforth, 1994; Baumeister et al., 1996). Hence, perpetrating bullying behavior and a high but unstable self-esteem should be associated, as proposed by Zapf and Einarsen (2003). Unstable self-esteem can be viewed in different directions, however. Does an unstable self-esteem indicate instability only across time, or is this instability reflected across measures and scale-items applied to map this psychological construct? The latter focus will be taken in the present study. There is a link between hostile or aggressive behavior and lack of social competency. For example, studies have portrayed that sex offenders have limited skills with respect to close relationships, having a hostile, unempathetic style of relating to others, particularly women (Hudson & Ward, 2000). Social and emotional competence requires the ability to detect, understand, and respond appropriately to the feelings of others (Frey, Hirschstein, & Guzzo, 2000). Social competency is closely linked with empathy, the capacity to share the emotional state of another, and is also associated with altruistic behavior (Eisenberg, 1986). Hudson and Ward (2000) contend that deficits in social competency, specifically those aspects relevant to close relationships, are clearly linked to engaging in offending or humiliating behavior against others. A low level of social competency also appears to be a dominant factor among many perpetrators of bullying, at least according to some anecdotal case stories (Adams, 1992). Lack of self-reflection and perspective taking may be important antecedent conditions to workplace bullying, which is why some people become bullies. According to anecdotal reports, perpetrators repeatedly report that they are not aware of the consequences of their behavior (Leymann, 1987). Social competency may be negatively related to anxiety or insecurity. Childhood studies have demonstrated that perpetrators usually report low levels of anxiety or insecurity, or they are roughly average in such dimensions (Olweus, 1991; Pulkkinen & Tremblay, 1992).

Bullying and Micropolitical Behavior A quite different explanatory model of bullying has been proposed, focusing on micropolitical behavior as an antecedent of workplace bullying behavior (Salin, 2003; Zapf & Einarsen, 2003). The concept of micropolitics is based on the premise that organizations do not consist of fully determined structures and processes (Neuberger, 1989). Thus, the organization expects its members to assist and close the gaps in the formal structure, for instance, by striving for achieving personal goals, participating in decision making, improving

738 138

Matthiesen and Einarsen

their level of influence, protecting their status, or rivalry. Intense micropolitical behavior may, however, be experienced as highly stressful for those involved. For example, it may result in increased levels of role conflict and role ambiguity because of frequent collisions between various role senders or role sets (van Sell, Brief, & Schuler, 1981). Thus, micropolitical behavior can lead to role stress. Role stress, such as role ambiguity and role conflict, can also be regarded as integrated parts of ongoing micropolitical behavior at the workplace. One may also argue that high levels of role stress are an invitation to micropolitical behavior. In line with this, emotional abuse at work is found to be associated with role conflict and role ambiguity (Keashly, Hunter, & Harvey, 1997). A function of micropolitical behavior could be that it is a strategy for protecting one’s self-interests and improving one’s own position (Zapf & Einarsen, 2003). Hence, micropolitical behavior may result in frequent episodes of interpersonal conflicts or aggression (De Roche, 1994). Bullying may therefore constitute an extreme type of micropolitical behavior or be a next-step consequence after intense interpersonal striving or conflicts, in line with the conflict-escalating model of Glasl (1980). If so, both perpetrators and targets of bullying should report higher levels of role stress and micropolitical behavior.

TARGETS OF BULLYING Several studies have investigated targets of bullying, searching for individual antecedent conditions or risk factors of victimization. A Norwegian study of psychiatric and personality disorders, using the MMPI-2 among 85 targets of bullying, portrayed clinically elevated levels of psychosomatic complaints, depressive thoughts, compulsory behavior, and paranoid or disturbed thoughts (Matthiesen & Einarsen, 2001). Zapf (1999), in a German study, identified two subgroups of bullying targets. One of the subgroups of bullied victims could not be distinguished from the control group in terms of personality dimensions measured. The other subgroup, however, tended to be significantly higher in anxiety and depression as well as lower in social skills. A forthcoming study (Glasø, Matthiesen, Nielsen, & Einarsen, in press), in which 79 bullied victims are compared with a matched control group, also revealed that bullied victims could be split into two subgroups, when the so-called big five personality perspective (McCrae & John, 1992) was applied. About 75% of the bullied group was not significantly different from the matched control group in terms of “big five” personality, whereas the other subgroup was significantly different on four out of five dimensions (among those, being higher on neuroticism and lower on agreeableness). These Norwegian and German studies highlight the notion that there may be different types of victims, with different pre-existing personality traits (such as general self-esteem and social competency). Some of the empirical findings regarding victim characteristics in schoolyard bullying may be important to note. Typically, victims tend to be more anxious and insecure than what is typical among the pupils in general (Lagerspetz, Björkquist, Berts, & King, 1982; Olweus, 1993). They are often cautious, sensitive, and quiet (Olweus, 2003). Olweus labels this victim group as passive or submissive. When bullied, they frequently react by crying and withdrawal. Thus, the targets of school bullying evidence low self-esteem (Carney & Merrell, 2001; Rigby & Slee, 1993). Most commonly, they have few friends as a source of emotional support, and often have higher rates of problems such as depression or anxiety (Carney & Merrell, 2001). Targets of bullying also have negative attitudes toward violence and violent means, according to Olweus (2003). Olweus concludes that

Bullying at Work 

739 139

the behavior of passive-submissive victims signals to others that they are insecure and worthless individuals, who will not retaliate if they are attacked or insulted.

Provocative Victims Olweus (1978), furthermore, labeled a subgroup of victims as “the provocative victims.” They are characterized by a combination of both anxious and aggressive reaction patterns. This group of children often has problems with concentration, and behaves in ways that may cause irritation and tension in their surroundings, be it among fellow pupils or among teachers. They also risk social isolation or exclusion, because others perceive their behavior as annoying and aggressive. Hence they may bully younger and weaker children while being bullied themselves by older and more powerful children. Such targets may have a history of involvement in bullying situations, both as targets and as perpetrators. Such bullying targets have also been studied among adults. A U.K. study of 5,288 adults in various workplace settings (Smith, Singer, Hoel, & Cooper, 2003) asked respondents to recall their experiences with childhood bullying. An association between childhood and workplace bullying was found: Former school victims were found to be more exposed to bullying at work. This childhood-adulthood link was especially pronounced among the bullying victims. Some 11% among former school victims and 13% among former school bully-victims reported that they were exposed to bullying at work. The corresponding numbers were found to be 9% for an ordinary student group as well for the former school bullies. Palmer and Thakordas (2005) made a study among 70 male imprisoned offenders. Here, the bully-victim group reported higher levels of hostility than the other offenders. In this selected prison sample, some 43% were categorized as belonging to the bully-victim group (12% were pure bullies and 16% were classified as pure victims).

AIMS OF THE STUDY The first aim of the present study was to investigate whether targets and perpetrators of bullying at work portray certain personality characteristics. Second, what is the relative number of provocative victims among self-reported targets of bullying? Provocative victims are defined as those employees who admit to having bullied others at the work place as well as claiming to be targets of bullying. The third aim of the study was to ascertain the number of self-reported perpetrators of bullying in a diverse sample of leaders and employees. The fourth aim was to investigate whether role stress and role ambiguity characterize workplaces where bullying flourishes, creating a fertile soil for intense micropolitical behavior. Micropolitical behavior, as measured by the concepts of role stress and role ambiguity, has been suggested as a major antecedent of bullying at work. We therefore suggest that bullying may be particularly common in work situations characterized by a high level of role stress. Unclear or conflicting demands and expectations around tasks, obligations, privileges, and priorities lay the foundation for a high level of micropolitical behavior among highly frustrated individuals, which will increase the risk of bullying behaviors. The following hypotheses will be investigated: 1. Provocative victims will report more prior acquaintance with bullying compared to others victims, be it in (a) former job(s) or (b) in their childhood. Provocative victims will also (c) report more childhood experiences as perpetrators of bullying.

740 140

Matthiesen and Einarsen

2. Perpetrators of bullying will report high levels of aggression, a high but unstable level of self-esteem, and a low level of social competence. 3. Provocative victims will (a) report a low level of self-esteem, combined with a high level of aggressiveness, and low level of social competence. Targets of bullying will report a low level of self-esteem combined with a low level of social competence. 4. Targets of bullying, as well as provocative victims and perpetrators, will report an elevated level of role conflict and role ambiguity.

METHOD Respondents Respondents participating in this cross-sectional survey study were randomly selected members from six Norwegian labor unions and members of the Norwegian Employers’ Federation (NHO). The participating unions, all situated in the geographical area around the city of Bergen, represented a convenient sample that reflected a diversity of work environments, hence increasing the validity and generalizability of the results. The labor unions included the Teachers Union, the Union of Hotel and Restaurant Workers, the Union of Trade and Commerce (consisting mainly of employees in shops and the administration of private businesses), the Union of Graphical Workers, the Union of Electricians, and the Union of Clericals and Officials (consisting of employees in the city administration of Bergen). All samples were studied as part of a more comprehensive research project on bullying and harassment in the workplace (see also Einarsen, Raknes, & Matthiesen, 1994; Einarsen, Raknes, Matthiesen, & Hellesøy, 1994; Einarsen & Skogstad, 1996). A total of 4,742 labor union members and employers’ representatives were selected from a total population of 10,616 individuals (a 47% response rate). The number of respondents (4,742) is equal to the number of valid questionnaires received. All of the questionnaires were distributed by mail. In the total sample 53% are men and 47% are women. All age groups between 16 and 70 are covered, with a mean age of 38 years (SD = 11.9). About 12% are 25 years of age or less, whereas 5% of the respondents are aged 60 or above. About 80% are employed on a full-time basis. Most of the respondents (62%) are employed in private enterprises. Thus, 38% are public employees. Furthermore, about 20% of the respondents work in organizations with more than 100 employees, and 9% in organizations with 5 or fewer employees.

Questionnaires The questionnaire used in the present study consisted of demographic variables, healthrelated variables, scales on psychological traits, single questions and scales on harassment and bullying, and scales and questions measuring perceived work environment quality. The following scales and measurements were included in the present study. Three single questions measured exposure to bullying at work during the last 6 months as well as earlier exposure to bullying (in present job, or in earlier jobs). Prior to these questions the respondents were presented with the following definition of bullying: “Bullying (harassment, badgering, niggling, freezing out, offending someone) is a problem in some workplaces and for some workers. To label something as bullying it has to occur repeatedly over a period of time, and the person confronted has to have difficulties defending

Bullying at Work 

741 141

himself/herself. It is not bullying if two parties of approximately equal ‘strength’ are in conflict or the incident is an isolated event.” The question on exposure to bullying and harassment was stated as follows: “Have you been subjected to bullying at the work place during the last six months?” The response categories were: “no,” “occasionally,” “now and then,” “about once a week,” and “many times a week.” All respondents who confirmed that they “occasionally” or more often were targets of bullying were defined as being exposed to bullying at work. This single question has been shown to be a valid measure of exposure to bullying at work (Mikkelsen & Einarsen, 2001; Salin, 2001). The respondents were also asked if they had bullied others at the workplace. Respondents who confirmed that this was the case were defined as perpetrators of bullying. Provocative victims are those who claim to be both a victim and a perpetrator of bullying. Two additional questions addressed childhood experiences with bullying at school, be it as a target or a perpetrator of bullying. Two questions addressed earlier experiences of bullying at work. In addition to the single questions on bullying, the Negative Acts Questionnaire (NAQ) and Bergen Bullying Index (Einarsen & Raknes, 1997; Einarsen, Raknes, & Matthiesen, 1994; Einarsen, Raknes, Matthiesen, et al., 1994) were administrated. The NAQ consists of 18 items measuring exposure to negative episodes or situations typical of bullying and may be regarded as a quantitative inventory on exposure to bullying, according to Einarsen and Raknes (1997) and Mikkelsen and Einarsen (2001). The response categories for NAQ are “daily,” “weekly,” “sometimes,” and “never.” The Cronbach’s alpha of NAQ was found to be 0.86, indicating a high internal stability. The Bergen Bullying Index is a global measurement of perceived individual and organizational consequences of bullying and harassment. It consists of five items, each scored on a 4-point Likert scale from “agree strongly” to “disagree strongly.” The Cronbach’s alpha was 0.82, indicating satisfactory internal stability. Personality Traits. The study contained three measures of aggressive tendencies adopted from research on schoolyard bullying (Olweus, 1987, 1991). Aggression after provocation was measured by three items (with Cronbach’s alpha = 0.76). Aggression against superiors contained two items (Cronbach’s alpha = 0.52). Aggression against peers was measured by three items (Cronbach’s alpha = 0.70). One measure on self-esteem was included in the study, measuring general self-esteem (Alsaker & Olweus, 1986; Rosenberg, 1965). The measure of general self-esteem, consisting of six items, had satisfactory internal stability (Cronbach’s alpha = 0.84). A four-item measure of social anxiety was also added (Alsaker & Olweus, 1986), measuring perceived incompetence and anxiety in social settings. Cronbach’s alpha for this scale was found to be 0.73. In total, the personality inventories consisted of 25 items, all formulated as statements describing oneself as a person. Six response alternatives were applied, with the range from agree completely (1) to disagree completely (6). Negative statements were reversed, resulting in a positive–negative continuum for all scales, with a theoretical range of 1–6 points, with 1 denoting the most positive value. Role Conflict and Role Ambiguity. These constructs were measured using two scales developed by Rizzo, House, and Lirzman (1970). Role conflict, consisting of eight items, measures the degree to which one perceives contradictory expectations, demands, or values in one’s job. Role ambiguity, consisting of six items, measures the degree to which the respondents experience their job situation as unpredictable and not clarified. Both scales are scored on a 7-point Likert scale, ranging from “totally agree” to “totally disagree.” Both scales are scored in a positive–negative direction, with 1 denoting the most positive value. Cronbach’s alphas were found to be 0.78 and 0.81 for role conflict and role ambiguity, respectively.

742 142

Matthiesen and Einarsen

Statistics. The data were analyzed by the use of SPSS 11.5 statistical package. The following analyses were performed: One-way ANOVA, reliability analysis with Cronbach’s alpha, cross-tab analysis for categorical data, and frequency statistics.

RESULTS Some 8% of the sample reported exposure to bullying at work (Figure 1). In addition, 2% can be defined as provocative victims. About 5% of the respondents admitted to being perpetrators of bullying at work. In sum, about 1 in 10 respondents reported to be a victim of bullying, whereas 1 in 20 acknowledged acting as a perpetrator of bullying. Respondents with no experience of bullying were used as a comparison group (n = 1,838). Most provocative victims (78%) and most perpetrators (74%) were males (χ2 = 34.32, df = 3, p < .001). Most targets and provocative victims (60% and 80%, respectively) were employed in private sector (χ2 = 21.80, df = 3, p < .001). Both present and former targets of bullying, as well as perpetrators, were overrepresented in companies with 100 or more employees (χ2 = 40.87, df = 15, p < .001), and in companies with skewed gender distribution of either males or females (χ2 = 31.62, df = 6, p < .001). Union representatives made up about 25% of both the targets and the perpetrator groups, whereas they constituted 36% of the provocative victim group (χ2 = 18.47, df = 3, p < .001). In the rest of the sample 18% were union representatives.

Acquaintance With Bullying The first hypothesis addressed whether provocative victims would report more prior experiences with bullying compared to other victims, be it in former jobs or in their childhood. In general, the provocative victim group reported more former workplace experience as bullying targets, relative to all other groups. Thirty-two percent of the provocative victims admitted that they had been bullied earlier in their career, in another workplace. The corresponding number in the target group is 17%, 10% in the perpetrator group, and 5% in the comparison group. The group differences are significant (χ2 = 77.54, df = 3, p < .001). The provocative victims also report more experienced bullying than do the others with respect to childhood experiences (Figure 2). Forty-eight percent of the provocative victims

Figure 1. Three bullying groups, percentage distribution (number of respondents in parentheses). The comparison group is omitted (84.3%, n = 1,838).

Bullying at Work 

743 143

claimed that they were bullied as children. In the target group, as well as among the bullies, the corresponding number was 27%, whereas 19% of the comparison group had experienced bullying in childhood (χ2 = 53.18, df = 9, p < .001). The provocative victims also frequently admit that they acted as bullies in their childhood. Forty-five percent in this group report childhood experiences as bullies. This number is higher than for the bullies (38%) and for the targets of bullying (14%). The group difference is significant (χ2 = 94.38, df = 9, p < .001). Summarized, we may regard the first hypothesis as confirmed.

Perpetrators Our second hypothesis stated that perpetrators will report a high level of aggression, high but unstable self-esteem, and low social competence. The results showed that perpetrators scored significantly higher on aggressiveness compared to targets, provocative victims, and the comparison group (see Table 1 for details). The perpetrators describe themselves as more aggressive after provocation than the others (M = 3.92, as compared to mean values between 3.43 and 3.77 for the other three groups; two out of three LSD post hoc tests were significant). Correspondingly, the perpetrator group reported a higher level of aggression against superiors (M = 3.58, as compared to mean values between 3.40 and 3.50 for the other three groups; two out of three post hoc tests were significant). On the other hand, provocative victims reported a higher level of aggression against peers than did the other groups, including the perpetrator group (M = 2.51, whereas the mean values were 2.24 for the perpetrators, and 1.92 and 1.91 for the bullied victims group and the comparison group, respectively). Reliability analyses using Cronbach’s alpha were conducted to investigate the second part of the second hypothesis, predicting that the perpetrators have a high but unstable within-scale self-esteem. The hypothesis can be verified to the extent that the perpetrators report a high level of self-esteem combined with a low internal stability (that is, a markedly lower level of Cronbach’s alpha than for the other three groups) on the scale measuring self-esteem. As shown in Table 1, perpetrators portray a higher level of self-esteem compared to the two target groups (lack of self esteem scores: M = 2.28 versus M = 2.67 for

Figure 2. for both).

Childhood bullying experiences. Group comparisons (χ2 = 53.19 and 94.39, p < .001

744 144

Matthiesen and Einarsen

the target group and M = 2.75 for the provocative victim group). The highest level of selfesteem was reported by the comparison group (lack of self-esteem score of M = 2.19). Unstable self-esteem can be reflected as low interitem reliability score, summarized as Cronbach’s alpha reliability coefficients. Whereas the Cronbach’s alpha for the general selfesteem scale was found to be 0.76 in the perpetrator group, the corresponding numbers for the other groups were 0.82 (targets), 0.69 (provocative victims), and 0.81 (the comparison group). Hence, the perpetrator group seems to have a more unstable self-esteem than the comparison group, as the internal stability of the scale measuring self-esteem is somewhat lower. The provocative victims, who are also perpetrators themselves, portrayed an internal stability lower than 0.70 (that is, 0.69) on general self-esteem. Coefficient 0.70 is seen as the criterion for a sufficient level of internal stability by Nunnally (1978). Higher instability was also verified, using a sum score procedure, in which all combinations of self-esteem items minus the rest of the other items, compared separately, was calculated. The item comparisons were then summed up to constitute a measure of variability in one’s response set. Thus, low score on the variability measure denotes high stability when responding to the items of the scale. Using this procedure, it was found that the comparison group had the lowest level of self-esteem instability (M = 23.54, SD = 7.79), whereas the provocative victim group revealed the highest level (M = 27.43, SD = 8.85). The scores for the bullying target group and the perpetrator group were in-between (respectively M = 25.12, SD = 9.18 and M = 24.66, SD = 7.50). The group difference regarding self-esteem instability was significant (one-way ANOVA; F (3/ 2064) = 5.56, p < .001). Post hoc tests (LSD procedure) revealed that provocative victims had a more unstable self-esteem than the comparison group ( p = .01), whereas the other group of bullies, the perpetrator group, did not differ. Summed up, the second hypothesis was partially verified. Only one of the perpetrator groups, the provocative victims, reported a higher self-esteem instability.

Targets and Provocative Victims The third hypothesis indicated that provocative victims will report a low level of self-esteem, combined with a high level of aggressiveness and a low level of social competence. The hypothesis also states that targets of bullying will report a lower level of self-esteem combined with a lower level of social competence. The provocative victims scored lower on selfesteem and social competency than did the target group and the comparison group (M = 2.75 and M = 2.67, as opposed to M = 2.67 and M = 2.53 for the bullying target group and M = 2.19 and M = 2.40 for the comparison group). Thus, the provocative victim group, as target group, reports a lower level of social self-esteem and social competency than the comparison group. Correspondingly, Table 1 portrays that the provocative victims report more aggression than do targets of bullying, and also more than the comparison group (the three measures of aggression vary between M = 3.77 and M = 2.51 for the provocative victims, targets scored between M = 3.43 and M = 1.92, while the comparison group scored between M = 3.63 and M = 1.91). Only the perpetrator group report more aggression, in terms of aggression after provocation and aggression against superiors, than the provocative victim group. Provocative victims did, however, report more aggression against their friends as compared to the perpetrator group. In sum, the third hypothesis can be regarded as supported.

Role Stress The fourth hypothesis stated that perpetrators, targets, and provocative victims will report an elevated level of role conflict and role ambiguity. High levels of such role stress may lead

TABLE 1.

Personality Differences Between Targets, Provocative Bullied Victims, Bullies, and a Comparison Group

Lack of self-esteem and social competency General self-esteem Social competency Aggression Aggression after provocation Aggression against superiors Aggression against friends

Targets (1)

Provocative Victims (2)

Perpetrators (3)

Comparison Group (4)

M (SD)

M (SD)

M (SD)

M (SD)

Post hoca

df

F

2.67 (1.05) 2.53 (0.85)

2.75 (0.91) 2.67 (0.68)

2.28 (.81) 2.46 (0.68)

2.19 (.85) 2.40 (0.72)

1> 3–4, 2> 3–4 1>4

3/ 2124 3/ 2130

21,14*** 3.21*

3.43 (1.22)

3.77 (1.30)

3.92 (1.14)

3.63 (1.21)

14

3/ 2130

3.94**

3.40 (0.62)

3.50 (0.76)

3.58 (0.68)

3.41 (0.63)

14

3/ 2124

3.03*

1.92 (0.82)

2.51 (1.22)

2.24 (0.79)

1.91 (0.81)

14, 3>4

3/ 2130

12.78***

Note. One-way ANOVA with post hoc tests. a The numbers in post hoc column refer to significant pair-wise group comparisons. LSD procedure. *p < .05. **p < .01. ***p < .001. Range: 1–4, in which 1 reflects the most positive value.

TABLE 2. Variations in Micropolitical Behavior, Operationalized as Group Differences in Role Conflict and Role Ambiguity: Group Comparison of Targets, Provocative Bullied Victims, Bullies, and a Comparison Group

Role conflict Role ambiguity

Targets (1)

Provocative Victims (2)

Perpetrator Group (3)

Comparison Group (4)

M (SD)

M (SD)

M (SD)

M (SD)

Post Hoc Testsa

df

F

3.68 (1.21) 3.01 (1.17)

3.88 (1.13) 3.24 (1.20)

3.25 (1.11) 2.72 (0.91)

2.81 (1.13) 2.56 (0.93)

1 > 3–4, 2 > 3–4, 1 >3–4, 2 > 3–4, 3 > 4

3/ 2132 3/ 2130

45.99* 18.58*

Note. One-way ANOVA with post hoc tests. numbers in post hoc column refer to significant pair-wise group comparisons. LSD procedure. *p < .001. Range: 1–7, 1 reflects the most positive value.

aThe

Bullying at Work 

747 147

to micropolitical behavior that in turn may be perceived as or may escalate into bullying, as aggression provoked by frustration, for example. Table 2 provides an overview of the group comparisons regarding role conflict and role ambiguity. One-way ANOVA statistics revealed significant group differences between the four subsamples ( p < .001 for both role conflict and role ambiguity), with the three bullying groups reporting elevated levels of role stress compared to the comparison group. Table 1 portrays that the two groups of bullying targets, respectively provocative victims and the other target group, reported a higher level of role conflict (M = 3.88 and M = 3.68) than the perpetrator group (M = 3.25) and the comparison group (M = 2.81). The group differences are much the same for role ambiguity (M = 3.01 and M = 3.24, as compared to M = 2.72 and M = 2.56). The perpetrators also experienced significantly more role stress in terms of role ambiguity than did the comparison group, according to the post hoc test. In sum, the fourth hypothesis may therefore be regarded as supported.

DISCUSSION The present study has shown that while some 8% claim to be targets of bullying at work, another 2% may be classified as provocative victims, a concept borrowed from schoolyard bullying studies (Olweus, 2003). Additionally, 4% claim to be perpetrators of bullying. Systematic exposure to negative social experiences in childhood, such as bullying, may cause serious negative aftereffects in adult life (e.g., a vulnerable personality). Thus, empirical findings indicate that persons who were victimized at school are more likely to be victimized in the workplace (Smith et al., 2003). The target group, and especially the provocative victims, has more prior experiences with bullying than others. They reported more exposure to bullying in the schoolyard but also in former jobs. For many of the provocative victims, bullying experience in school also consisted of experiences as perpetrators. Nearly one in two had such experiences. The notion addressed in our first hypothesis was thus confirmed. Some of the provocative victims may possess what Levinson (1978) labeled as an “abrasive personality.” An abrasive personality reflects a tendency to behave in an insensitive, ruthless way, especially when the person is confronted with social pressure situations. This study did also reveal an elevated level of aggressiveness among the provocative victims. Brodsky (1976) has pointed this out quite categorically, claiming that “after studying harassers and studying their victims, it seemed that there was never a victim who would not have made an excellent harasser” (p. 109). This statement seems to be supported in the present study as far as the provocative victims are concerned. Among schoolchildren, Lorber, Felton, and Reid (1984) found that victims of abuse are more often likely to be disruptive, aggressive, and violent than their nonabused counterparts. They explained such results as socially learned behaviors. A high level of self-esteem has generally been regarded as desirable and adaptive and has been used as an indicator of good adjustment (Taylor, 1989; Taylor & Brown, 1988). Schoolyard bullying research has indicated that the child bully in general seems to have high self-esteem (Olweus, 2003). In line with Olweus’s notion, our findings portrayed that the adult perpetrator report a relatively high level of self-esteem. However, an Irish study of over 8,000 schoolchildren showed that children who had been involved in bullying as victims, were bullies, or both had significantly lower global self-esteem than did children who had neither bullied nor been bullied (O’Moore & Kirkham, 2001). The notion that perpetrators are characterized by elevated levels of self-esteem runs contrary to an entrenched

748 148

Matthiesen and Einarsen

body of wisdom that has long pointed to low self-esteem as the root of violence and other destructive kinds of social behavior (Baumeister et al., 1996). Building on a theoretical assumption by Baumeister et al. that violent perpetrators are characterized by an elevated but unstable self-esteem, we proposed that perpetrators of workplace bullying will report an unstable self-esteem. This statement was only partially supported, however. The perpetrators did report a somewhat lower level of self-esteem stability, operationalized as high interitem correlations between the self-esteem measure items, compared to victims and nonvictims in the comparison group, although above the recommended threshold for acceptable internal stability recommended by Nunnally (1978). However, a low stability, below the recommended threshold, was found among the provocative victims. Furthermore, a measure of variability in response set between items showed significant differences between the groups, where provocative targets, followed by victims, showed the highest variability in scores, again indicating unstable self-esteem. Thus, the second hypothesis in the present study was supported, but for the provocative victims only. However, it is possible that our applied six-item measure of general self-esteem is not adequate or sensitive enough as a measure of stability-instability in self-esteem. A longitudinal design would probably be better, looking at stability over time. The perpetrators, comprising 5% of the sample, did report more aggressiveness than the victim group and the comparison group. The difference is unambiguous as perpetrators admit stronger aggressive reactions after provocations in the workplace. Correspondingly, the perpetrators also aggress more against superiors and against peers. This is in line with Olweus’s (1994, 2003) findings from school bullying research. The bullies in the workplace tend to react more aggressively than others across different social situations. This may lend support to an aggressiveness hypothesis, that is: The perpetrators in general manifest a more aggressive behavioral pattern at the workplace than others. This aggressiveness has been described by Ashforth (1994) as “petty tyranny” among superiors. The workplace, and the bullying that take place therein, may be one of several modes and arenas for the expression of aggressiveness in general. In line with this finding, bullying behaviors have been found to be strongly related to high levels of overall aggression, physical and verbal aggression, hostility, and anger (Buss & Perry, 1992). Our findings indicate that all targets of bullying, but especially the provocative victims, report a lower level of self-esteem and social competency than do perpetrators and the comparison group, in line with our third hypothesis. Research among schoolchildren has also found that provocative victims had the lowest self-esteem of the subgroups investigated (O’Moore & Kirkham, 2001). A possible explanation of this group difference may be that many victims possess a more fragile self-esteem, in accordance with Brockner’s (1988) “plasticity hypothesis,” an explanation also in line with our results on their unstable self-esteem. According to Brockner, individuals with so-called plastic self-esteem are very dependent on good work performance and positive appraisal from others to maintain their self-esteem. Hence, a low self-esteem may then easily become a consequence of bullying. The necessary self-esteem nurturance, leading to a stable, favorable global evaluation of oneself, thus cannot be obtained. It may also be that some of the targets of bullying feel frustrated or irritated because of lack of self-esteem support in the worksite, with the result that they act in a provocative manner, which others may perceive as a general lack of social competency. The low internal stability among provocative targets on the scale measuring self-esteem supports the hypothesis that these victims suffer from a fragile self-esteem. The importance of social competency in relation to bullying is also demonstrated by Coyne et al. (2003) in their Irish workplace study. Self- and peer-reported victims tended

Bullying at Work 

749 149

to be more likely than controls to have difficulty coping with personal criticism, to be easily upset, to view the world as threatening, to be anxious, tense, and suspicious of others. Among schoolchildren, social intelligence has been found to be negatively related to victimization (Kaukainen et al., 2002). However, bullying has also been explained by references to factors in the psychosocial work environment (Leymann, 1996; Einarsen, 2000). For instance, bullying has been claimed to be a consequence of micropolitical behavior, rivalry, and competition in the worksite (Zapf et al., 2003). The compression of career structures resulting from delayering processes represents fewer opportunities for advancement in many workplaces, thereby increasing the competition between managers for promotion to a shrinking pool of jobs (Sheehan, 1999), with growing interpersonal conflict and bullying as possible outcomes (Hoel & Salin, 2003). Role stress, that is, work roles that interfere with each other (role conflict) or work roles that are experienced as unclear or confusing (role ambiguity) may indicate a certain need for micropolitical behavior. The fourth hypothesis in the present study addressed this issue, claiming that the perpetrator group in particular, with their self-reported bullying behavior, but also the two groups of bullying targets, would report this kind of job stress. The aforementioned hypothesis was supported. The perpetrator group portrayed a higher level of role stress indicative of a need for micropolitical behavior than did the comparison group. However, targets of bullying, both the provocative victims as the well as the targets only, reported an even higher level of experienced role stress than did the perpetrators. Previous findings have shown that bullying seems to occur in stressful, competitive, and negative working environments (Vartia, 1996), in particular in workplaces with a high level of role conflict. Thus, it should come as no surprise that targets of bullying report exposure to micropolitical behavior reflected by role conflict and role ambiguity, equal to or even higher than levels experienced by the perpetrator group.

Methodological Constraints The present study was conducted by the use of survey method. The respondents were asked about sensitive topics, issues that may have led them to respond with denial or social desirability. The respondents were asked not only if they felt subjected to bullying at the workplace, but also if they had acted as perpetrators themselves. Hence, an individual research level was applied (Matthiesen, Aasen, Holst, Wie, & Einarsen, 2003). The result may be valid to the extent that the respondents answered the questions in an honest way and in accordance with their inner, subjective experience. However, measurement of perpetrator behavior is methodologically difficult. The bullying definition presented in our study, and most other bullying studies, does not require any intention about bullying among the perpetrators to meet the definition. When someone is asked if they have bullied others, many will reject or deny the question, because they may never have had any intent to do so. Still, the other part in a dyadic relationship at the worksite may feel subjected to bullying. To admit perpetrator behavior in retrospect, thus, requires a certain level of empathy or skill to imagine social situations from the perspective of other people. “False negative” bullies or provocative victims may then pose a methodological problem in the present study. “False positives” may of course also exist, self-critical individuals who are eager to admit perpetrator behavior, but without ever being in the power position or acting in a manner where others actually feel subjected to bullying. Social desirability (Crowne & Marlowe, 1964) or guilt reduction (Sigmund Freud, see, e.g., Gomez, 2005) may cause perpetrators of bullying to reject that they act in the role of a bully.

750 150

Matthiesen and Einarsen

Another limitation inherent within the present study is its cross-sectional design. The general self-esteem among the targets of bullying, and the proneness to express frustration or complex aggression overtly or covertly, may constitute a function of former bullying exposure, to give an example. The behavior pattern of the victims may have changed as a consequence of bullying. It is possible that some individuals possess a higher propensity to become troubled with bullying than most others, because they have a more vulnerable personality (Matthiesen & Einarsen, 2001). Negative perceptions of the work situation have previously been found to be associated with workplace deviance, with personality traits such as conscientiousness, emotional stability, and agreeableness as moderating factors (Colbert, Mount, Harter, Witt, & Barrick, 2004).

CONCLUSION This study has portrayed, in line with previous research on bullying among children, that individual differences exist when perpetrators and targets of bullying are compared. Bullied victims can be divided into at least two groups: targets of bullying and provocative victims. The provocative victim group has only briefly been focused on in previous workplace research and deserves more attention. Correspondingly, it is time to shed more light on the perpetrator in workplace bullying, and not only from the perspective of the target. Only a few papers have been published so far, such as the aforementioned survey study carried out by Coyne et al. (2003), where perpetrators and their responses have been mapped. Most studies on workplace bullying are anecdotal in this respect, with the consequence that perpetrators mostly have been studied at a distance. This study may therefore provide an important contribution to the field by breaking the “at a distance” research habit regarding research on perpetrators of workplace bullying.

REFERENCES Adams, A. (1992). Bullying at work—How to confront and overcome it. London: Virago. Alsaker, F. D., & Olweus, D. (1986). Assessment of global negative self-evaluations and perceived stability of self in Norwegian pre-adolescents and adolescents. Journal of Early Adolescence, 6, 269–278. Ashforth, B. (1994). Petty tyranny in organizations. Human Relations, 47, 755–778. Baumeister, R. F., Smart, L., & Boden, J. M. (1996). Relation of threatened egotism to violence and aggression: The dark side of high self esteem. Psychological Review, 103, 5–33. Beck, A., Rush, A., Shaw, B., & Emery, G. (1979). Cognitive theory of depression. New York: Guilford Press. Bing, S. (1992). Crazy bosses: Spotting them, serving them, surviving them. New York: William Marrow. Brockner, J. (1988). Self-esteem at work. Lexington, MA: Lexington Books. Brodsky, C. M. (1976). The harassed worker. Toronto: Lexington Books. Buss, A. H., & Perry, M. (1992). The aggression questionnaire. Journal of Personality and Social Psychology, 63, 452–459. Carney, A. G., & Merrell, K. W. (2001). Bullying in schools: Perspectives on understanding and preventing an international problem. School Psychology International, 22(3), 364–382. Colbert, A. E., Mount, M. K., Harter, J. K., Witt, L. A., & Barrick, M. R. (2004). Interactive effects of personality and perception of the work situation on workplace deviance. Journal of Applied Psychology, 89(4), 599–609.

Bullying at Work 

751 151

Coyne, I., Chong, P. S.-L., Seigne, E., & Randall, P. (2003). Self and peer nominations of bullying: An analysis of incident rates, individual differences, and perceptions of the working environment. European Journal of Work and Organizational Psychology, 12(3), 209–228. Crowne, D., & Marlowe, D. (1964). The approval motive: Studies in evaluative dependence. New York: John Wiley. De Roche, C. P. (1994). On the edge of regionalization: Management style and the construction of conflict in organizational change. Human Organization, 53(3), 209–219. Einarsen, S. (2000). Harassment and bullying at work: A review of the Scandinavian approach. Aggression and Violent Behavior, 4(5), 379–401. Einarsen, S., Hoel, H., Zapf, D., & Cooper, C. L. (2003). The concept of bullying at work: The European tradition. In S. Einarsen, H. Hoel, D. Zapf, & C. L. Cooper (Eds.), Bullying and emotional abuse in the workplace (pp. 3–30). London: Taylor & Francis. Einarsen, S., & Raknes, B. I. (1997). Harassment in the workplace and the victimization of men. Violence and Victims, 12, 247–263. Einarsen, S., Raknes, B. I., & Matthiesen, S. B. (1994). Bullying and harassment at work and their relationships to work environment quality—An exploratory study. European Work and Organizational Psychologist, 4(4), 381–401. Einarsen, S., Raknes, B. I., Matthiesen, S. B., & Hellesøy, O. H. (1994). Mobbing og harde personkonflikter. Helsefarlig samspill på arbeidsplassen [Bullying and tough interpersonal conflicts. Health injurious interaction at the workplace]. Bergen: Sigma Forlag. Einarsen, S., & Skogstad, A. (1996). Bullying at work: Epidemiological findings in public and private organizations. European Journal of Work and Organizational Psychology, 5(2), 185–201. Eisenberg, N. (1986). Altruistic emotion, cognition and behavior. Hillsdale, NJ: Erlbaum. Forsyth, D. R. (2006). Group dynamics (4th ed.). Belmont, CA: Thomson Wadsworth. Frey, K. S., Hirschstein, M. K., & Guzzo, B. A. (2000). Second step: Preventing aggression by promoting social competence. Journal of Emotional and Behavioral Disorders, 8(2), 102–112. Glasl, F. (1980). Konfliktmanagement: Diagnose und Behandlung von Konflikten in Organizationen. Bern: Haupt. Glasø, L., Matthiesen, S. B., Nielsen, M. B., & Einarsen, S. (in press). Do targets of workplace bullying portray a general victim personality profile? Scandinavian Journal of Psychology. Gomez, L. (2005). The Freud wars: An introduction to the philosophy of psychoanalysis. London: Routledge. Griffin, R. S., & Gross, A. M. (2004). Childhood bullying: Current empirical findings and future directions for research. Aggression and Violent Behavior, 9, 379–400. Hoel, H., & Cooper, C. L. (2000). Destructive conflict and bullying at work. Manchester, UK: Manchester School of Management. Hoel, H., & Salin, D. (2003). Organizational antecedents of workplace bullying. In S. Einarsen, H. Hoel, D. Zapf, & C. L. Cooper (Eds.), Bullying and emotional abuse in the workplace (pp. 203–218). London: Taylor & Francis. Hudson, S. M., & Ward, T. (2000). Interpersonal competency in sex offenders. Behavior Modification, 24(4), 494–527. Kaukainen, A., Salmivalli, C., Lagerspetz, K., Tamminen, M., Vaura, M., Maki, H., et al. (2002). Learning difficulties, social intelligence, and self-concept: Connections to bully-victim problems. Scandinavian Journal of Psychology, 43, 269–278. Keashly, L., Hunter, S., & Harvey, S. (1997). Abusive interaction and role state stressors: Relative impact on student residence assistant stress and work attitudes. Work and Stress, 11, 175–185. Keashly, L., & Jagatic, K. (2003). By another name: American perspectives on workplace bullying. In S. Einarsen, H. Hoel, D. Zapf, & C. L. Cooper (Eds.), Bullying and emotional abuse in the workplace (pp. 31–61). London: Taylor & Francis. Kernis, M. H., Cornell, D. P., Sun, C. R., Berry, A., & Harlow, T. (1993). There is more to selfesteem than whether it is high or low: The importance of stability of self-esteem. Journal of Personality and Social Psychology, 65, 1190–1204.

752 152

Matthiesen and Einarsen

Kile, S. (1990). Helsefarlige ledere og medarbeidere [Health dangerous leaders and their co-workers]. Oslo: Hjemmets bokforlag. Lagerspetz, K. M., Björkquist, K., Berts, M., & King, E. (1982). Group aggression among school children in three schools. Scandinavian Journal of Psychology, 23, 45–52. Levinson, H. (1978). The abrasive personality. Harvard Business Review, 56, 86–94. Leymann, H. (1987). Mobbing i arbeidslivet [Bullying in working life]. Oslo: Friundervisningens forlag. Leymann, H. (1990). Mobbing and psychological terror at workplaces. Violence and Victims, 5(2), 119–126. Leymann, H. (1992). Psykiatriska halsoproblem i samband med vuxenmobbning. En rikstackende undersøkning med 2428 intervjuer [Psychiatric problems after mobbing—a study of 2428 individuals] (Report no. 3). Stockholm: Arbetarskyddsstyrelsen. Leymann, H. (1996). The content and development of bullying at work. European Journal of Work and Organizational Psychology, 51, 165–184. Leymann, H., & Gustavson, A. (1996). Mobbing at work and the development of post-traumatic stress disorders. European Journal of Work and Organizational Psychology, 5(2), 251–275. Lorber, R., Felton, D. K., & Reid, J. B. (1984). A social learning approach to the reduction of coercive process in child abuse families: A molecular analysis. Advances in Behavior Research and Therapy, 6, 29–45. Matthiesen, S. B., Aasen, E., Holst, G., Wie, K., & Einarsen, S. (2003). The escalation of conflict: A case study of bullying at work. International Journal of Management and Decision Making, 4(1), 96–112. Matthiesen, S. B., & Einarsen, S. (2001). MMPI-2 configurations after persistent bullying at work. European Journal of Work and Organizational Psychology, 10(4), 467–484. Matthiesen, S. B., & Einarsen, S. (2004). Psychiatric distress and symptoms of PTSD after bullying at work. British Journal of Guidance and Counseling, 32(3), 335–356. Matthiesen, S. B., Raknes, B. I., & Rokkum, O. (1989). Mobbing pa arbeidsplassen [Bullying at the worksite]. Journal of the Norwegian Psychological Association, 26(11), 761–774. McCrae, R. R., & John, O. P. (1992). An introduction to the Five-Factor model and its applications. Journal of Personality, 60(2), 175–215. Mikkelsen, E. G., & Einarsen, S. (2001). Bullying in Danish work-life: Prevalence and health correlates. European Journal of Work and Organizational Psychology, 10(4), 393–413. Mikkelsen, E. G., & Einarsen, S. (2002a). Basic assumptions and post-traumatic stress among victims of workplace bullying. European Journal of Work and Organizational Psychology, 11(1), 87–111. Mikkelsen, E. G., & Einarsen, S. (2002b). Relationships between exposure to bullying at work and psychological and psychosomatic health complaints: The role of state negative affectivity and generalized self-efficacy. Scandinavian Journal of Psychology, 43(5), 397–405. Neuberger, O. (1989). Mikropolitik als Gegenstand der Personalentwicklung [Micropolitics as object of personnel development]. Zeitschrift fuer Arbeits und Organisationspsychologie, 33(1), 40–46. Niedl, K. (1995). Mobbing/bullying am Arbeitsplatz. Munchen: Rainer Hampp Verlag. Niedl, K. (1996). Mobbing and well-being: Economic and personnel development implications. European Journal of Work and Organizational Psychology, 5, 203–214. Nielsen, M. B., Matthiesen, S. B., & Einarsen, S. (2005). Ledelse og personkonflikter: Symtomer på postraumatisk stress blant ofre for mobbing fra ledere [Leadership and bullying. Posttraumatic symptoms among victims after bullying by their leaders]. Nordisk Psykologi, 57(4), 391–415. Nunnally, J. O. (1978). Psychometric theory. New York: McGraw-Hill. O’Moore, M., & Kirkham, C. (2001). Self-esteem and its relationship to bullying behaviour. Aggressive Behavior, 27, 269–283. Olweus, D. (1978). Aggression in the schools: Bullies and whipping boys. Washington, DC: Hemisphere, Wiley. Olweus, D. (1987). Bully/victim problems among schoolchildren in Scandinavia. In J. P. Mykebust & R. Ommundsen (Eds.), Psykologprofesjonen mot år 2000 [The psychological profession towards year 2000] (pp. 395–413). Oslo: Universitetsforlaget.

Bullying at Work 

753 153

Olweus, D. (1991). Bully/victim problems among schoolchildren: Basic facts and effects of a school based intervention program. In K. Rubin & D. Pepler (Eds.), The development and treatment of childhood aggression (pp. 411–448). Hillsdale, NJ: Lawrence Erlbaum. Olweus, D. (1993). Bullying at school. What we know and what we can do. Oxford: Blackwell Publishers. Olweus, D. (1994). Bullying at school: Basic fact and effect of a school based intervention program. Journal of Child Psychology and Psychiatry, 35, 1171–1190. Olweus, D. (2003). Bully/victim problems in school. Basic facts and an effective intervention programme. In S. Einarsen, H. Hoel, D. Zapf, & C. L. Cooper (Eds.), Bullying and emotional abuse in the workplace (pp. 62–78). London: Taylor & Francis. Palmer, E. J., & Thakordas, V. (2005). Relationship between bullying and scores on the Buss-Perry aggression questionnaire among imprisoned male offenders. Aggressive Behavior, 31, 55–66. Pulkkinen, L., & Tremblay, R. E. (1992). Pattern of boys’ social adjustment in two cultures and at different ages: A longitudinal perspective. International Journal of Behaviour Development, 15, 527–553. Rayner, C., & Cooper, C. L. (2003). The black hole in “bullying at work research.” International Journal of Management and Decision Making, 4(1), 47–64. Rigby, K., & Slee, P. T. (1993). Dimensions of interpersonal relation among Australian children and implications for psychological well-being. Journal of Social Psychology, 133(1), 33–42. Rizzo, J. R., House, R. J., & Lirzman, S. I. (1970). Role conflict and ambiguity in complex organizations. Administrative Science Quarterly, 15, 150–163. Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press. Salin, D. (2001). Prevalence and forms of bullying among business professionals: A comparison of two different strategies for measuring bullying. European Journal of Work and Organizational Psychology, 10(4), 425–441. Salin, D. (2003). Ways of explaining workplace bullying: A review of enabling, motivating and precipitating structures and processes in the work environment. Human Relations, 56(10), 1213–1232. Sheehan, M. (1999). Workplace bullying: Responding with some emotional intelligence. International Journal of Manpower, 20(1–2), 57–69. Smith, P. K., Singer, M., Hoel, H., & Cooper, C. L. (2003). Victimization in the school and the workplace: Are there any links? British Journal of Psychology, 94, 175–188. Taylor, S. E. (1989). Positive illusions: Creative self-deception and the healthy mind. New York: Basic Books. Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: A social psychological perspective on mental health. Psychological Bulletin, 103, 193–210. van Sell, M., Brief, A. P., & Schuler, R. S. (1981). Role conflict and role ambiguity: Integration of the literature and directions for future research. Human Relations, 34, 43–71. Vartia, M. (1996). The sources of bullying—Psychological work environment and organizational climate. European Journal of Work and Organizational Psychology, 5, 203–205. Zapf, D. (1999). Organisational, work group related and personal causes of mobbing/bullying at work. International Journal of Manpower, 20(1/2), 70–85. Zapf, D., & Einarsen, S. (2003). Individual antecedents of bullying. In S. Einarsen, H. Hoel, D. Zapf, & C. L. Cooper (Eds.), Bullying and emotional abuse in the workplace (pp. 165–184). London: Taylor & Francis. Zapf, D., Einarsen, S., Hoel, H., & Vartia, M. (2003). Empirical findings on bullying in the workplace. In S. Einarsen, H. Hoel, D. Zapf, & C. L. Cooper (Eds.), Bullying and emotional abuse in the workplace (pp. 103–126). London: Taylor & Francis. Correspondence regarding this article should be directed to Professor Stig Berge Matthiesen, PhD, University of Bergen, Faculty of Psychology, Psychosocial Science Unit, Christiesgate 12, N-5015 Bergen, Norway. E-mail: [email protected]

Violence and Victims, Volume 27, Number 3, 2012

Workplace Bullying, Emotions, and Outcomes Lars Glasø, PhD BI Norwegian Business School, Oslo, Norway University of Bergen, Norway

Guy Notelaers, PhD School of Business and Economics, Department of Organization and Strategy, Maastricht University, The Netherlands University of Bergen, Norway This study examines emotional experiences as potential mediators between exposure to workplace bullying and job satisfaction, organizational commitment, and intention to leave the organization, respectively. A total of 5,520 respondents participated in the study. Drawing upon affective events theory (AET), the results show that emotions partly mediate these relationships and, hence, support the notion that emotions play a central part in the relationship between bullying and essential occupational outcomes.

Keywords: bullying; emotions; affective events theory; job satisfaction; organizational commitment; intention to leave the organization

B

ullying at work has been recognized as a serious problem in contemporary working life and has been defined as harassing, offending or socially excluding someone, or negatively affecting someone’s work tasks. Although the negative and unwanted nature of the behavior involved is essential to the concept of bullying, the core characteristic of bullying is not necessarily the type of the behaviors involved per se but rather the pattern and persistency of these experiences (Einarsen & Hoel, 2008). For example, being ignored by a manager or colleagues at work may happen to anyone, but if repeated over a long period of time, such relatively harmless behaviors may be experienced as acts of bullying. Thus, bullying is normally not about single or isolated events but rather about behaviors that are repeated and persistently directed at one or more individuals. Approximately 5%–30% of the European workforce is found to be exposed to some kind of bullying behavior (Nielsen et al., 2009), and one out of four victims leave their job as a result of such experiences (Rayner, 1997). Consequences of such exposure may be both psychologically and physically debilitating for the targets (Balducci, Alfano, & Fraccaroli, 2009; Glasø, Nielsen, Einarsen, Haugland, & Matthiesen, 2009; Hogh, Mikkelsen, & Hansen, 2011; Rodríguez-Muñoz, Baillien, De Witte, Moreno-Jiménez, & Pastor, 2009). Bullying may also take its toll on relationships and family life and has been associated with organizational costs such as turnover (Rayner, 1997), absenteeism (Kivimäki, Elovainio, & Vatera, 2000), and decreased organizational commitment and

© 2012 Springer Publishing Company

155

156

Glasø and Notelaers

productivity (Hoel, Einarsen, & Cooper, 2003). Accordingly, bullying can be costly for the individual, the organization, and the society. However, despite the large body of evidence pointing to such harmful consequences of bullying, little is known about factors pertaining to how and why bullying may produce such outcomes in targets. From a stress-theoretical point of view (e.g., Lazarus, 1999; Lazarus & Cohen-Charash, 2001; Lazarus & Folkman, 1984), emotions may be of great importance to understand the outcomes of stressful transactions such as workplace bullying. In accordance with this idea, emotions are essential to understanding people’s reactions at work (Muchinsky, 2000). Research within the field of victimology (e.g., Janoff-Bulman, 1992) has shown that facing harassment at work may be experienced as a trauma and, as such, may generate severe emotional reactions such as fear, anxiety, and shock, as well as lower the target’s state of positive feelings. Qualitative studies have shown that emotions such as anxiety, fear, anger, helplessness, and irritability may follow the experience of being bullied (Ayoko, Callan, & Härtel, 2003). Furthermore, in an Irish interview study, 30 targets of bullying reported high levels of anxiety and depression (O’Moore, Seigne, McGuire, & Smith, 1998), whereas targets at a university in Great Britain reported a high prevalence of shame (Lewis, 2004). Leymann (1990) has also documented high levels of depression, helplessness, anger, anxiety, and despair among bullied targets in psychiatric clinics and such feelings seem to be representative for targets regardless of gender, position, and age (Tracy, Lutgen-Sandvik, & Alberts, 2006). Positive emotions also seem to be related to workplace bullying in terms of decreased intensity (see Bowling & Beehr, 2006; Glasø, Løkke Vie, Holmdal, & Einarsen, 2011). In line with this view, a study focusing on relationships between stressful events and positive and negative moods experienced by white-collar workers showed that exposure to the stressors intensified their negative emotional experiences while reducing the intensity of their positive mood (van Eck, Nicolson, & Berkhof, 1998). Recently, Brotheridge and Lee (2010) have demonstrated a significant reduction of happiness among targets of workplace bullying compared to nonvictims. Even though both negative and positive emotions certainly are connected to bullying, most studies mention affective states solely as consequences or end products of bullying (see meta-analysis by Bowling & Beehr, 2006), thereby leaving out the possible mediating effect emotions may have on the targets’ attitude and subsequent behavior. Nevertheless, a few studies examining affective experiences as possible mediators between workplace bullying and different outcomes do exist. For example, Penhaligon, Louis, and Restubog (2009) tested the mediating role of perceived rejection between workgroup mistreatment and affective outcomes such as depression and organizationbased self-esteem among 142 part-time work participants. The results indicated that perceived rejection mediates the relationship between mistreatment and the outcomes. A study among 224 Danish factory employees found that negative affect partly mediated the relationship between exposure to bullying and self-reported health complaints (Mikkelsen & Einarsen, 2002). However, the picture is blurred; in a Norwegian study, Matthiesen and Einarsen (2004) did not find any significant mediator or moderator effects of positive or negative emotions between bullying and health problems. Moreover, Glasø et al. (2011), who investigated personality traits and experienced emotions among targets, found that the relationships between exposure to bullying and both job satisfaction and intention to leave were actually partly mediated by the targets’ emotional experiences. These studies indicate that emotional reactions among bullied targets may mediate different outcome variables. As such, the results are mixed and emphasized the need for more

Emotions and Workplace Bullying

157

studies examining emotional experiences among targets of workplace bullying. Hence, to learn more about mechanisms that may be involved in the development of individual reactions associated with bullying, the present study examines the role of positive and negative emotions as potential mediators between exposure to workplace bullying and outcomes such as organizational commitment, job satisfaction, and intention to leave the organization in a large and heterogeneous sample. Affective events theory (AET; Weiss & Cropanzano, 1996) elucidates what happens between work events and subsequent employee attitudes and behaviors by focusing on the role of emotions. AET offers a broad description of the causes, consequences, and structure of affective experiences at work. According to AET, work environment features (i.e., roles and job designs) influence attitudes directly through a cognitive route as well as indirectly through an affective route, the latter by determining the occurrence of positive or negative affective work events. Experiencing such “hassles” and “uplifts” at work lead to negative and positive affective reactions, which in turn lead to affect-driven behaviors and work attitudes. Work attitudes may then influence judgment-driven behavior, such as turnover intentions. Weiss and Cropanzano (1996) claim that workplace events activate affective responses, which after being accumulated over time will influence workplace attitudes such as job satisfaction. The model is based on the assumption that emotions are not equal to job satisfaction. More specifically, job satisfaction, rather than constituting an affective phenomenon, is conceptualized as an evaluative judgment of or attitude toward one’s job (Spector, 1997). Such an attitude should therefore not be confused with genuine emotions that employees experience at work, because emotions have causes and consequences that are distinguishable from the causes of evaluative judgement such as job satisfaction (Weiss & Cropanzano, 1996). Hence, in the present study we focus on accumulated emotions by measuring the respondents’ positive and negative emotional reactions during the preceding 2 weeks. Although AET is regarded as a significant contribution toward explaining the causes and consequences of emotions at work, we still need more empirical examination of the basic assumptions put forward in the model (Briner & Totterdell, 2002; Glasø et al., 2011; Weiss & Beal, 2005). The model does not specify the type of work environments or work events that may be associated with positive or negative affective reactions. According to Basch and Fisher (2000), few studies have explored the specific events described in AET that might arouse affect at work. In the present study, a certain kind of negative work event is examined, namely the experience of bullying behaviors at work. Thus, our aim is to investigate whether AET’s predictions regarding the fundamental role of emotions can be corroborated by the relationships between bullying and some occupational outcomes, which are described in the following section. Many studies have shown that exposure to workplace bullying is associated with lowered job satisfaction (Baruch, 2005; Bilgel, Aytac, & Bayram, 2006; Hauge, Skogstad et al., 2010; Hubert, Furda, & Steenma, 2001; Quine, 2003), an association that was recently confirmed in a longitudinal study (Rodríguez-Muñoz et al., 2009). In addition, different cross-sectional studies have reported that exposure to workplace bullying is related to low levels of organizational commitment (Hubert et al., 2001; Notelaers & De Witte, 2003). Several studies have also shown that exposure to bullying is associated with high turnover intention among targets (Baruch, 2005; Hauge et al., 2010; LutgenSandvik, 2006; Rayner, 1997). In their meta-analysis focusing on antecedents and consequences of workplace harassment, Bowling and Beehr (2006) found that environmental and emotional factors were significantly related to the targets’ well-being. These find-

158

Glasø and Notelaers

ings fit the predictions suggested in AET, namely that negative events (i e., exposure to bullying) will affect the recipients’ work attitudes such as job satisfaction, organizational commitment, or intention to leave the organization. However, none of these studies examined the central premise of AET, namely to what extent individuals’ accumulated emotional responses to workplace events mediate the relationship between such events and their cognition and behavior (see Weiss & Beal, 2005), which is the main focus of the present study. In accordance with AET, we hypothesize the following: Hypothesis 1: Negative emotions mediate the relationships between exposure to bullying and job satisfaction, organizational commitment, and turnover intention, respectively. Specifically, exposure to bullying will be associated with an increase in negative emotions, which in turn will be related to a decrease in job satisfaction and organizational commitment and an increased intention to leave the organization, respectively. Hypothesis 2: Positive emotions mediate the relationships between exposure to bullying and job satisfaction, organizational commitment, and turnover intention, respectively. Specifically, exposure to bullying will be associated with a decrease in positive emotions, which in turn will be related to a decrease in job satisfaction and organizational commitment and an increased intention to leave the organization, respectively.

METHOD Sample Data was collected from a large survey of psychosocial risk factors at work during the period 2003–2007 by means of an anonymous self-report questionnaire and distributed to 12 organizations in Belgium who wanted to conduct a psychosocial risk analysis. Questions about workplace bullying appeared at the very end of the questionnaire. Participation was voluntary, and anonymity was guaranteed by the researchers at the Directorate of the Research of Working Conditions. The organizations operated in the manufacturing industry (25%), the service sector (44%), and the public sector (31%). With respect to size, one organization was small (fewer than 100 employees), three organizations had between 100 and 250 employees, and the remaining eight organizations employed more than 500 people. A total of 5,520 respondents completed the survey, yielding a response rate of 70%, which is well above the mean found in surveys of this kind (see Baruch & Holtom, 2008). A small majority of the respondents were male, whereas 42% were female. The mean age of the respondents was 40.5 years (SD 5 10). Approximately 8% were blue-collar workers, 40% were white-collar workers, 32% were public servants, 18% held managerial positions, and the remaining respondents held social welfare or health care functions (nurses, social workers, medical doctors). Approximately 80% had a permanent contract and worked full-time. Fifty-three percent of the respondents spoke Dutch, whereas 47% spoke French. Although this is not a representative sample for the Belgian workforce, it is a heterogeneous sample that allows us to investigate the current mediation hypotheses.

Instruments Exposure to bullying behaviors at work was measured by the short version of the negative acts questionnaire (Einarsen, Hoel, & Notelaers, 2009; Notelaers & Einarsen, 2008). This scale presents respondents with nine behavioral items referring to both direct

Emotions and Workplace Bullying

159

(e.g., verbal abuse, offensive remarks) and indirect (e.g., social isolation, slander) negative acts. All items were described in behavioral terms with no reference to the phrase “bullying,” thus, measuring perceived exposure to bullying behaviors without forcing the respondents to label these situations as bullying. For each item, the respondents were asked how often they had been exposed to the specific behavior during the preceding 6 months. The response categories were “never,” “now and then,” “about monthly,” “weekly and more often.” In the present study, the scale showed a satisfactory internal consistency measured by Cronbach’s alpha (a 5 .82). Emotional experiences were measured by the Questionnaire on the Evaluation and Experience of Work (QEEW; van Veldhoven & Meijman, 1994). This scale consists of two scales referring to six negative emotions (a 5 .84) and six positive emotions (a 5 .88). The respondents were asked to what extent they generally felt the listed emotions during the preceding 2 weeks. The following emotions were listed in the questionnaire: nervous, optimistic, gloomy, at ease, dejected, calm, agitated, sad, relaxed, uncomfortable, cheerful, and elated. Response categories were “not at all,” “hardly,” “somewhat,” and “­completely.” All three outcomes were measured with items included in the QEEW (see also Notelaers, De Witte, van Veldhoven, & Vermunt, 2007; van Veldhoven & Meijman, 1994). Job satisfaction was measured with five items. Examples of items are “I do my work because I have to” (reversed), and “Mostly, I am pleased to start on my day’s work.” Cronbach’s alpha 5 .79. Organizational commitment was measured with five items, and this scale had a reliability of .82. Examples of items are “It is important to me that I can make a contribution to the organization’s business,” and “I really feel very closely involved with this organization.” Turnover intention was measured with four items, such as “I sometimes think about changing my job,” and “Next year, I plan to look for a job outside this organization.” This scale also showed satisfactory Cronbach’s alpha values (a 5 .78).

Analyses and Statistics According to AET, emotional reactions are considered to mediate the relationships between exposure to workplace bullying and the three outcome variables. To test a mediation model with more than three variables, the traditional Baron and Kenny (1986) approach is less appropriate (MacKinnon, Lockwood, Hoffman, West, & Sheets, 2002; McQueen, Getz, & Bray, 2003). Structural equation modeling, however, allows for analyzing several mediators and several outcome variables simultaneously, thereby taking measurement error into account as well. Because the positive and negative emotions were highly correlated in the measurement model, the issues of multicollinearity and suppression recommend modeling of the positive and the negative emotions separately (Maassen & Bakker, 2001). In the present study, we follow an analytical strategy employed by Geurts, Kompier, Roxburgh, and Houtman (2003) and distinguish between three models to assess mediation. In the first model, we assume that emotions fully mediate the relation between exposure to bullying and the three outcome variables. In this model, job satisfaction, turnover intention, and organizational commitment are directly explained by emotions. In the second model, we assume that the relation between workplace bullying and the three outcome variables is partially mediated. In this model, job satisfaction, organizational commitment, and turnover intention are explained by emotions and exposure to workplace bullying. If emotions are still contributing to the explained variance of the outcome variables, we can conclude that emotions partially mediate the relationship between workplace bullying and the outcomes. To evaluate the extent of mediation, a third model is estimated. This model

160

Glasø and Notelaers

differs from the second model because the path coefficients between emotions and three outcome variables are fixed. These parameters are set equal to the estimated parameters of the first model, in which complete mediation was estimated. The difference in x2 between the second and third model indicates whether emotions mediate the relation between workplace bullying and the three outcome variables. If the difference in x2 between the second and third models is not significant, the model assuming complete mediation is the most appropriate one. If the difference in x2 between the second and third models is significant, the model in which partial mediation is assumed is the most suitable one. And finally, when the difference in x2 between the second and third models is significant and the relation between emotions and outcome variables is not significant in the second model, the model assuming no mediation is the most suitable one. These three models are further described based on various commonly reported statistical criteria. We used goodness-of-fit indices; the root mean square error of approximation (RMSEA), the non-normed fit index (NNFI), the comparative fit index (CFI), and the standardized root mean square residual (SRMR). RMSEAs and SRMRs that are smaller than .08 are indicative of a satisfactory approximate fit of the theoretical model (Hu & Bentler, 1999). For the other indices such as the NNFI and the CFI, values greater than .90 (and preferably greater than .95) are considered to indicate a good fit.

RESULTS Table 1 shows some essential descriptive statistics and the Pearson’s correlation coefficients for the variables in the study. All correlation coefficients between the variables were significantly different from zero (p , .001). The means in the table show that, on

TABLE 1.  Means, Standard Deviations, and Correlations Between Latent Variables Means and Standard Job Intention Organizational Workplace Deviations Emotions Satisfaction to Leave Commitment Bullying Emotions

1.94 (0.68)

2,99 (0.72)

.44

2.51

.54

Job  satisfaction

0.81 (0.29)

.61

2.62

.82

2.49

Intention to  leave

0.29 (0.33)

2.41

2.65

2.67

.40

Organizational  commitment

0.67 (0.31)

.53

.82

2.67

Workplace  bullying

1,38 (0.41)

2.47

2.53

.43

2.59 1

1

1 2.49

.46 1

Note. All correlations were significant at the 0.001 level. Above diagonal 5 correlations between negative emotions and the study’s variables; below diagonal 5 correlations between positive emotions and the study’s variables.

Emotions and Workplace Bullying

161

TABLE 2.  Goodness-of-Fit Indices and Model Comparisons for Negative Emotions x2 (d.f.)

Dx2(Ddf)

p

RMSEA

NNFI

CFI

SRMR

0.00

.062

.94

.95

.082

Model 1

6323.47 (373)

Model 2

6287.01 (370)

M1 vs. M2 36.46 (3)

0.001

.062

.94

.95

.081

Model 3

6304.59 (373)

M2 vs. M3 17.58 (2)

0.001

.061

.94

.95

.079

average, the respondents in this sample are modestly exposed to workplace bullying, show somewhat more negative than positive emotions, are very highly satisfied with their current job, are highly committed to the organization, and show a low intention of turnover. Table 2 contains the fit measures of the LISREL analysis (Jöreskog & Sörbom, 1993) for the first model (which assumed complete mediation by negative emotions), as well as the fit measures for models where emotions partially mediate the relation between workplace bullying and the outcome variables job satisfaction, organizational commitment, and turnover intention. The first model or basic model (Model 1) in which complete mediation was assumed fitted the data reasonably well. However, with 373 degrees of freedom, the x2 of 6323.47 is too high to obtain a perfect fit. Because the sample size is very large, the x2 is highly inflated. The RMSEA, a measure of approximate fit and other descriptive statistics, is therefore also used. The descriptive statistics showed that this model had a satisfactory fit. In this model, the entire path coefficients between negative emotions and workplace bullying and between negative emotions and outcome variables are significant (t . 1.96). Hence, negative emotions may be conceived as mediators. Model 2 is the model in which negative emotions partially mediate the relationship between exposure to workplace bullying and the outcome variables. This partial mediation model did lead to a significant improvement of fit (Dx2(3) 5 36.46). The partial mediation model also fitted the data reasonably well. To test the extent of the mediation, we conceptualized Model 3 where the path coefficients between negative emotions and outcome variables were fixed to the estimates in the full mediation model. As shown in Table 2, this model leads to a significant deterioration of fit (Dx2(1) 5 17.58). Hence, negative emotions do not fully mediate the relationships between exposure to workplace bullying and job satisfaction, organizational commitment or turnover intentions, respectively. However, because the negative emotions were significant predictors of these three outcome variables in the second model, we may conclude that negative emotions partially mediate the relationships between exposure to workplace bullying and job satisfaction, organizational commitment and turnover intention, respectively. The total amount of explained variance for job satisfaction was 35%, 13% for turnover intention, and 27% for organizational ­commitment. As shown in Figure 1, workplace bullying was positively related to negative emotions (b 5 0.48, p , .001). All paths between emotions and the outcome variables were significant. When a simple regression analysis was employed where no mediation was modeled,

Job satisfaction

1 Naq 2

Js 1

.70

Js 2

.65

.08

Naq 3

.55

Naq 4 .69 .09

Workplace bullying

.62

Turnover intention

1

Ti 1

.89

Ti 2

.40 .46

Naq 6

.59 .67

.48

.32

Naq 7 .57 Negative emotions

.60

.45

Organizational commitment

1

Oc 1

.44

Oc 2

.65

.56 Erw 1

.79 Erw 3

Figure 1.  Mediation model of negative emotions.

.81 Erw 5

.46

.71

.75 Erw 7

Erw 8

Oc 3

.68 Oc 4 Oc 5

Erw 10

Glasø and Notelaers

Naq 9 1

Ti 3 Ti 4

.11

Naq 8

Js 4 Js 5

.60

Naq 5

Js 3

.78 .62

.69

162

Naq 1

1

Emotions and Workplace Bullying

163

TABLE 3.  Goodness-of-Fit Indices and Model Comparisons for Positive Emotions x2 (d.f.)

Dx2(Ddf)

P

RMSEA

NNFI

CFI

SRMR

0.00

.064

.94

.95

.079

Model 1

6836.47 (373)

Model 2

6782.93 (370)

M1 vs. M2 53.54 (3)

0.001

.064

.94

.95

.077

Model 3

6798.80 (373)

M2 vs. M3 15.82 (2)

0.001

.064

.94

.95

.074

the direct relationship between workplace bullying and job satisfaction expressed as a standardized path coefficient (b) was 20.43. In the partial mediation model, this coefficient dropped to 20.08. Regarding the relationship between workplace bullying and turnover intention, the beta in the regression model dropped from 0.34 to 0.09 in the final model. Correspondingly, the relationship between workplace bullying and organizational commitment revealed a drop from b 5 20.41 to b 5 20.11. Hence, the initial relationships were decimated, indicating that negative emotions may act as substantial mediators between workplace bullying and all three outcomes. This result indicates that our first hypothesis is partly supported. In examining the potential mediator role of positive emotions, we followed the same procedure and employed the same analyses as for the negative emotions (see previous ­discussion). Table 3, which contains the fit statistics and the model comparison procedure to evaluate the potential mediation effects, shows that also positive emotions partially mediate the relationships under investigation. The path coefficients of the partial mediation model are shown in Figure 2. The total amount of explained variance for job satisfaction was 41%, 15% for turnover intention, and 36% for organizational commitment. As expected, workplace bullying was negatively related to positive emotions (b 5 20.39, p , .001). All paths between emotions and the outcome variables were significant. Workplace bullying was still weakly related to turnover intention and to job satisfaction. When a simple regression analysis was estimated where no mediation was modeled, the direct relationship between workplace bullying and job satisfaction expressed as a standardized path coefficient (b) was 20.44. In the partial mediation model, the corresponding path coefficient dropped to 20.11. Whereas the relationship between workplace bullying and turnover intention was 0.34 in the regression model, it dropped to 0.11 in the final model. The relationship between workplace bullying and organizational commitment dropped from 20.42 to 20.11. Hence, the initial relationships decreased substantially, indicating that positive emotions also act as a substantial mediator between workplace bullying and all three outcomes. This result indicates that our second hypothesis is partly supported. An essential principle of the AET model is that job satisfaction, organizational commitment, and emotions are related but clearly distinguishable constructs. The results of the present study lend only partial support to this view, because both job satisfaction and positive and negative emotions (r 5 0.64, p , .01; r 5 20.55, p , .01, respectively) as

Job satisfaction

1 Naq 2

Js 1

.70

Js 2

.65

.11

Naq 3

.46

Naq 4 .69 .11

Workplace bullying

.62

Turnover intention

1

Ti 1

.89

Ti 2

.40 .46

Naq 6

.59 .67

.39

.38

Naq 7 .57 Positive emotions

.60

Organizational commitment

.60

1

Oc 1

.44

Oc 2

.65

.56 Erw 2

.69 Erw 4

Figure 2.  Mediation model of positive emotions.

.54 Erw 6

.66

.84

.82 Erw 9

Erw 11

Oc 3

.68 Oc 4 Oc 5

Erw 12

Glasø and Notelaers

Naq 9 1

Ti 3 Ti 4

.11

Naq 8

Js 4 Js 5

.60

Naq 5

Js 3

.78 .62

.69

164

Naq 1

1

Emotions and Workplace Bullying

165

well as organizational commitment and positive and negative emotions(r 5 0.60, p , .01; r 5 20.45, p , .01, respectively) were strongly correlated. Although AET’s contribution to the field seems important, it draws heavily on existing theories of emotion and has not been comprehensively tested (Briner & Totterdell, 2002). Hence, we wanted to explore this issue further by comparing different factor models. The analyses showed that the fivefactor model distinguishing between the five factors in the current study obtained a better fit (negative emotions, x2 5 5276.83, df 5 367; positive emotions, x2 5 6053.51, df 5 367) than other factor models where, for instance, emotions and job satisfaction were conceived as one factor (negative emotions, x2 5 10936.91, df 5 371; positive emotions, x2 5 10914.96, df 5 371) or a factor model where emotions, job satisfaction, and organizational commitment were combined into one latent variable (negative emotions, x2 5 13356.02, df 5 374; positive emotions, x2 513061.33, df 5 374) and, finally, where next to workplace bullying only one factor was differentiated (negative emotions x2 5 15683.33, df 5 367; positive emotions x2 5 15655.46, df 5 376). Thus, our findings support the idea of AET that job satisfaction, organizational commitment, and turnover intentions and ­emotions are distinct empirical and theoretical constructs.

DISCUSSION The present study indicates that the relationships between exposure to bullying and job satisfaction, organizational commitment, and intention to leave the organization are partly but substantially mediated by the targets’ negative and positive emotional experiences. Hence, both our hypotheses are to a certain extent supported. AET points to affective experiences as being crucial in connection with the outcomes of work events. However, the present findings revealed strong but still only partial mediation. This is in line with several studies demonstrating the detrimental consequences of workplace bullying, and emphasizes the fact that exposure to persistent negative acts at work is felt as a deeply degrading experience with negative effects on both the victim’s self, identity, and health (see Glasø et al., 2009; Hogh et al., 2011; Leymann, 1996). A core problem of victimization because of workplace bullying is that such events may threaten or even shatter the target’s basic assumptions of being a valuable and competent person living in a safe and caring environment (Janoff-Bulman, 1992). Accordingly, it seems reasonable that the targets in the present study speak of strong negative emotional reactions as well as reduced levels of positive emotions. To explain the mediating effects of emotional reactions in this study, it may be fruitful to think of negative emotions as states producing a more critical attitude within individuals than do positive emotions (e.g., George, 2000). Hence, it seems sensible, and in accordance with AET, that the targets in the present study will—after having repeatedly experienced negative emotions during the preceding 2 weeks—become increasingly more attentive, doubtful of or skeptical toward their work environment, and that their emotional reactions as such may influence their attitudes regarding their job satisfaction, organizational commitment, and intention to leave the organization. Nevertheless, positive emotions measured in the present study such as optimism, cheerfulness, and calmness also mediated the relationship between bullying and the outcomes. Hence, bullying not only generates negative emotions, but it also seems to reduce the positive ones, and in effect reduces job satisfaction and organizational commitment while increasing the targets’ intentions to seek employment elsewhere. This result seems

166

Glasø and Notelaers

to support our second hypothesis as well as predictions of AET. However, this result is not in accordance with another central theory of emotions, namely the two-domain theory of emotions (TDT; Fisher, 2002), which claims that an individual’s emotional system consists of two separate parts, positive and negative affects, which are activated by different stimuli. Consistent with TDT, it could be reasoned that bullying is not likely to be related to the presence or absence of positive affect in targets because positive emotions are likely to be predicted by events other than bullying. On the other hand, appraisal theories of emotion (Lazarus, 1999; Roseman, 1991) claim that an individual’s appraisal of the situation determines his or her subsequently affective response. Being exposed to workplace bullying may undermine the target’s self confidence and, as shown in the present study, lead to a reduction of positive emotions within the target (see also Brotheridge & Lee, 2010). Furthermore, positive emotions may have significant behavioral implications concerning the targets’ coping efforts. For instance, Tugade, Fredrickson, and Barrett (2004) have shown that positive emotions play a crucial role in enhancing coping resources in the face of negative events. As shown in the present study, exposure to workplace bullying seems to decrease the intensity of the targets’ positive emotions and may accordingly have weakened their coping resources. In this respect, their emotional reactions may have influenced their coping capacity and, subsequently, their choice of coping strategy. If so, one such coping strategy may be the targets’ intentions to leave the organization because this act removes them from the source of the problem and accordingly reduces the emotional pain. Some other targets may quit in despair or because of sickness resulting from prolonged strain and stress. The mediating effect of emotions on the targets’ levels of job satisfaction, organizational commitment, and intention to leave may have serious negative organizational effects. Although research on organizational commitment among targets of workplace bullying is scarce, it is reasonable to expect a negative impact on productivity and performance where reduced commitment or withdrawal is used as a coping strategy (see also Hoel, Sheehan, Cooper, & Einarsen, 2011, for a similar view). In contrast, several studies have shown that employees with a strong organizational commitment are loyal to the organization and exhibit extra efforts at work (Meyer, Stanley, Herscovitch, & Toplonytsky, 2002), are highly motivated (Kacmar, Carlson, & Brymer, 1999), and showing less need to change their jobs (Jaros, 1997) than employees with a weak organizational commitment. In a British study, targets of bullying rated their own performance to be around 85% of normal capacity, whereas people with no bullying experiences reported 92% capacity (Hoel, Cooper, & Faragher, 2001). Moreover, the mediating effects of decreased levels of experienced positive emotions and increased levels of negative emotions shown in the present study may also influence the targets’ level of job engagement, creativity, and innovation (see Isen, 2000; Isen, Daubman, & Nowicki, 1987; Rayner, Hoel, & Cooper, 2002), which should be studied in more details in future studies. In a study on bullying and absenteeism in the United Kingdom, Hoel and Cooper (2000) found that victims of bullying took on average 7 days more sick leave per year than those who were neither bullied nor had witnessed that bullying had taken place. Based on a prevalence rate of bullying of 10%, this would account for a total of 18 million lost working days annually in the United Kingdom alone. Based on a meta-analysis of bullying research, Sheehan, McCarthy, Barker, and Henderson (2001) calculated a bullying cost in the order of AU$ 0.6–3.6 million per annum for an Australian business with 1,000 employees. Similarly, at a national level, Hoel et al. (2003) concluded that costs related to absence and replace-

Emotions and Workplace Bullying

167

ment because of bullying alone may account for close to £2 billion annually. However, one should bear in mind that intention to leave the organization as a research variable used in the present study is considered only to measure an individual’s attitude toward the job, as O’Connell, Calvert and Watson (2007) reported that 60% of some Irish targets considered leaving, whereas 15% actually left the organization. Nevertheless, bullying represents a considerable cost both to employees, employers, and the society (Di Martino, Hoel, & Cooper, 2003), and the results of the present study indicate that experienced emotions seem to play a central role in this process. Of course, a cognitive route may also partially account for these relationships and should be investigated in further studies. Finally, a crucial principle of the AET model is that emotion, job satisfaction, organizational commitment, and turnover intention are related but nevertheless dissimilar constructs. Our findings, however, only partially give support to this view, because these variables were strongly correlated. We therefore explored this issue further by comparing different factor models (see “Results” section). The analyses showed that the fivefactor model distinguishing between the five constructs examined in the present study obtained a better fit than other factor models. Thus, our findings support the idea of AET, namely that emotion, job satisfaction, organizational commitment, and turnover intention are distinct empirical and theoretical constructs and should therefore be explicated and studied as such (see also Ashkanasy, Zerbe, & Härtel, 2002, for a similar view). In this respect, we believe that the use of general job satisfaction or general organizational commitment measures as the only indicators of affective experiences at work has limited value.

METHODOLOGICAL ISSUES One important limitation of the present study stems from the use of single-source selfreported data. Although exposure of bullying, emotional experiences, job satisfaction, organizational commitment, and intentions to leave must be measured by self-reports, common method variance may have enhanced the overall strength of correlations (see Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). By measuring emotions, we chose emotional reactions from the QEEW because of its good psychometric qualities and the fact that this instrument is extensively used in the European Lowlands. However, because of multicollinearity (r . 0.65) and possible suppression effects, the different kinds of positive and negative emotions could not be investigated simultaneously. Therefore, future studies may use the Positive and Negative Affect Schedule (PANAS), which comprises two distinct dimensions (positive affect/negative affect), which is a psychometrically sound and widely used instrument in the field (see Watson, Clark, & Tellegen, 1988; Weiss & Cropanzano, 1996). Asking respondents to which extent they experience positive and negative emotions with a time frame of 2 weeks seems quite suitable considering the fact that AET emphasizes the point of measuring accumulated emotions in explaining the outcome variables examined in the present study. Furthermore, when bullying becomes more or less an enduring state, it may both act as a daily hassle and constitute a more permanent feature of the working environment. Following AET, bullying should then have both direct and indirect effects on the outcomes, in line with the findings of the present study. However, one should keep in mind that because of the cross-sectional design of the present study, we cannot conclude about the causal relationship between the variables.

168

Glasø and Notelaers

This may imply that reduced job satisfaction and reduced organizational commitment as well as plans to quit one’s job contribute to those emotional reactions reported in this study. Furthermore, such withdrawal may be understood from the perspective of ostracism (see Williams, 1997), which might explain the expelling process of the targets from the workplace. Only a longitudinal study would provide valid data about causality of the impact of emotions on the variables in question. It should also be noted that measuring emotions is complicated, because emotional experiences are variable and transient (see e.g., Ben-Ze’ev, 2000) and may be difficult to recall and report accurately long after they have occurred (Fisher, 2002). However, in this study we focused on accumulated emotions, which, according to AET, are central to explaining behavioral and attitudinal outcomes at work. Nevertheless, we think future research also should examine such relationships longitudinally on an hourly or daily basis.

CONCLUSION The present study has documented, in accordance with AET, that both negative and positive emotions partly mediate the relationships between exposure to bullying and outcomes such as job satisfaction, organizational commitment, and intention to leave the organization. Hence, the consequences of workplace bullying are severe, and the targets’ ­experienced emotions seem to be strongly connected to this problem.

REFERENCES Ashkanasy, N. M., Zerbe, W. J., & Härtel, C. E. J. (2002). Managing emotions in the workplace. New York: M. E. Sharpe. Ayoko, O. B., Callan, V. J., & Härtel, C. E., J. (2003). Workplace conflict, bullying, and counterproductive behaviours. The International Journal of Organizational Analysis, 11(4), 283–301. Balducci, C., Alfano V., & Fraccaroli, F. (2009). Relationships between mobbing at work and MMPI-2 personality profile, posttraumatic stress symptoms, and suicidal ideation and behavior. Violence and Victims, 24(1), 52–67. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51(6), 1173–1182. Baruch, Y. (2005). Bullying on the net: Adverse behavior on e-mail and its impact. Information & Management, 42, 361–371. Baruch, Y., & Holtom, B. C. (2008). Survey response rate levels and trends in organizational research. Human Relations, 61(8), 1139–1160. Basch, J., & Fisher, C. D. (2000). Affective events–emotion matrix: A classification of work events and associated emotions. In N. M. Ashkanasy, C. E. J. Härtel, & W. J. Zerbe (Eds.), Emotions in the workplace: Research, theory and practice (pp. 36–48). Westport, CT: Quorum Books. Ben-Ze’ev, A. (2000). The subtlety of emotions. Cambridge, MA: MIT Press. Bilgel, N., Aytac, S., & Bayram, N. (2006). Bullying in Turkish white-collar workers. Occupational Medicine, 56(4), 226–231. Bowling, N. A., & Beehr, T. A. (2006). Workplace harassment from the victim’s perspective: A theoretical model and meta-analysis. The Journal of Applied Psychology, 91(5), 998–1012. Briner, R. B., & Totterdell, P. (2002). The experience, expression and management of emotion at work. In P. Warr (Ed.), Psychology at work (5th ed., pp. 229–252). London: Penguin Books.

Emotions and Workplace Bullying

169

Brotheridge, C. M., & Lee, R. T. (2010). Restless and confused. Emotional responses to workplace bullying in men and women. Career Development International, 15(7), 687–707. Di Martino, V., Hoel, H., & Cooper, C. L. (2003). Preventing violence and harassment in the workplace. Luxembourg: Office for Official Publications of the European Communities. Einarsen, S., & Hoel, H. (2008). Bullying and mistreatment at work: How managers may prevent and manage such problems. In A. Kinder, R. Hughes, & C. L. Cooper (Eds.), Employee well-being support: A workplace resource (pp. 161–173). New York: Wiley. Einarsen, S., Hoel, H., & Notelaers, G. (2009). Measuring exposure to bullying and harassment at work: Validity, factor structure and psychometric properties of the Negative Acts Questionnaire– Revised. Work & Stress, 23(1), 24–44. Fisher, C. D. (2002). Antecedents and consequences of real-time affective reactions at work. Motivation and Emotion, 26(1), 3–30. George, J. M. (2000). Emotions and leadership: The role of emotional intelligence. Human Relations, 53(8), 1027–1055. Geurts, S. A. E., Kompier, M. A. J., Roxburgh, S., & Houtman, I. L. D. (2003). Does work-home interference mediate the relationship between workload and well-being? Evaluating model fit. In R. H. Hoyle (Ed.), Structural equation modelling: Concepts, issues and applications (pp. 76–99). Thousand Oaks, CA: Sage. Glasø, L., Løkke Vie, T., Holmdal, G., & Einarsen, S. (2011). An application of affective events theory to workplace bullying: The role of emotions, trait anxiety and trait anger. European Psychologist, 16, advance articles. Glasø, L., Nielsen, M., Einarsen, S., Haugland, K., & Matthiesen, S. (2009). A study of basic life assumptions and post-traumatic stress disorder among victims of workplace bullying. Journal of the Norwegian Psychological Association, 46, 153–160. Hauge, L. J., Skogstad A., et al. (2010). The relative impact of workplace bullying as a social stressor at work. Scandinavian Journal of Psychology, Early online. Hoel, H., & Cooper, C. L. (2000). Destructive conflict and bullying at work: Report. Manchester, United Kingdom: Manchester School of Management. Hoel, H., Cooper, C., L., & Faragher, B. (2001). The experience of bullying in Great Britain: The impact of organizational status. European Journal of Work and Organizational Psychology, 10(4), 443–465. Hoel, H., Einarsen, S., & Cooper, C. L. (2003). Organisational effects of bullying. In S. Einarsen, H. Hoel, D. Zapf, & C. L. Cooper (Eds.), Bullying and emotional abuse in the workplace: International perspectives in research and practice (pp. 145–162). London: Taylor & Francis. Hoel, H., Sheehan, M., Cooper, C. L., & Einarsen, S. (2011). Organisational effects of workplace bullying. In S. Einarsen, H. Hoel, D. Zapf, & C. L. Cooper (Eds.), Bullying and harassment in the workplace: Developments in theory, research, and practice (pp.129–148). Boca Raton, FL: CRC Press. Hogh, A., Mikkelsen, E. G., & Hansen, Å. M. (2011). Individual consequences of workplace ­bullying/mobbing. In S. Einarsen, H. Hoel, D. Zapf, & C. L. Cooper (Eds.), Bullying and harassment in the workplace: Developments in theory, research, and practice (pp. 107–128). Boca Raton, FL: CRC Press. Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indices in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. Hubert, A., Furda, J., & Steenma, H. (2001). Mobbing, systematisch pestgedrag in organisaties. Gedrag en Organisatie, 14(6), 378–395. Isen, A. M. (2000). Positive affect and decision making. In M. Lewis & J. M. Haviland-Jones (Eds.), Handbook of emotions (2nd ed., pp. 417–435). New York: Guilford Press. Isen, A. M., Daubman, K. A., & Nowicki, G. P. (1987). Positive affect facilitates creative problem solving. Journal of Personality and Social Psychology, 52(6), 1122–1131.

170

Glasø and Notelaers

Janoff-Bulman, R. (1992). Shattered assumptions: Towards a new psychology of trauma. New York: Free Press. Jaros, S. J. (1997). An assessment of Meyer and Allen’s (1991) three-component model of organizational commitment and turnover intentions. Journal of Vocational Behavior, 51(3), 319–337. Jöreskog, K. G., & Sörbom, D. (1993). LISREL 8: Structural equation modeling with the SIMPLIS command language. Chicago: Scientific Software International. Kacmar, K. M., Carlson, D. S., & Brymer, R. A. (1999). Antecedents and consequences of organizational commitment: A comparison of two scales. Educational and Psychological Measurement, 59(6), 976–994. Kivimäki, M., Elovainio, M., & Vatera, J. (2000). Workplace bullying and sickness absence in ­hosptial staff. Occupational and Environmental Medicine, 57(10), 656–660. Lazarus, R. S. (1999). Stress and emotion: A new synthesis. New York: Springer Publishing. Lazarus, R., & Cohen-Charash, Y. (2001). Discrete emotions in organizational life. In R. L. Payne & C. L. Cooper (Eds.), Emotions at work: Theory, research and applications for management (pp. 45–81). Chichester, United Kingdom: Wiley. Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal and coping. New York: Springer Publishing. Lewis, D. (2004). Bullying at work: The impact of shame among university and college lecturers. British Journal of Guidance and Counselling, 32(3), 281–299. Leymann, H. (1990). Mobbing and psychological terror at workplaces. Violence and Victims, 5(2), 119–126. Leymann, H. (1996). The content and development of mobbing at work. European Journal of Work and Organizational Psychology, 5(2), 165–184. Lutgen-Sandvik, P. (2006). Take this job and . . . : Quitting and other forms of resistance to workplace bullying. Communication Monographs, 73(4), 406–433. Maassen, G., & Bakker, A. (2001). Suppressor variables in path models: Definitions and interpretations. Sociological Methods & Research, 30(2), 241–270. MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets, V. (2002). A comparison of methods to test mediation and other intervening variable effects. Psychological Methods, 7(1), 83–104. Matthiesen, S. B., & Einarsen, S. (2004). Psychiatric distress and symptoms of PTSD among victims of bullying at work. British Journal of Guidance & Counselling, 32(3), 335–356. McQueen, A., Getz, J. G., & Bray, J. H. (2003). Acculturation, substance use, and deviant behavior: Examining separation and family conflict as mediators. Child Development, 74(6), 1737–1750. Meyer, J. P., Stanley, D. J., Herscovitch, L., & Topolnytsky, L. (2002). Affective, continuance, and normative commitment to the organization: A meta-analysis of the antecedents, correlates, and consequences. Journal of Vocational Behavior, 61(1), 20–52. Mikkelsen, E. G., & Einarsen, S. (2002). Bullying in Danish work-life: Prevalence and health correlates. European Journal of Work and Organizational Psychology, 10, 393–413. Muchinsky, P. M. (2000). Emotions in the workplace: The neglect of organizational behaviour. Journal of Organizational Behavior, 21, 801–805. Nielsen, M. B., Skogstad, A., Matthiesen, S. B., Glasø, L., Aasland, M. S., Notelaers, G., et al. (2009). Prevalence of workplace bullying in Norway: Comparisons across time and estimation methods. European Journal of Work and Organizational Psychology, 18(1), 81–101. Notelaers, G., & De Witte, H. (2003). De relatie tussen werkstress en pesten op het werk en welbevinden op het werk [The relationship between workstressors, workplace bullying and wellbeing]. Arbeidsmarktonderzoekersdag, 2003 Vrij Universiteit Brussel. W. Herremans. Reeks, De Arbeidsmarkt in Vlaanderen. Leuven, Steunpunt Werkgelegenheid, Arbeid en Vorming 26 mei 2003, VUB.: 139–163. Notelaers, G., De Witte, H., van Veldhoven, M., & Vermunt, J. K. (2007). Construction and validation of the short inventory to monitor psychosocial hazards. Médecine du Travail & Ergonomie, 44, 11–17.

Emotions and Workplace Bullying

171

Notelaers, G., & Einarsen, S. (2008). The construction and validity of the Short-Negative Acts Questionnaire. Paper presented at the 6th International Conference on Workplace Bullying: Sharing our Knowledge, Montreal, Canada. O’Connell, P. J., Calvert, E., & Watson, D. (2007). Bullying in the workplace: Survey report. Dublin, Ireland: Department of Enterprise Trade and Employment, Economic and Social Research Institute. O’Moore, M., Seigne, E., McGuire, L., & Smith, M. (1998).Victims of workplace bullying in Ireland. The Irish Journal of Psychology, 19(2–3), 345–357. Penhaligon, N. L., Louis, W. R., & Restubog, S. L. D. (2009). Emotional anguish at work; The ­mediating role of perceived rejection on workgroup mistreatment and affective outcomes. Journal of Occupational Health Psychology, 14(1), 34–45. Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. (2003). Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology, 88(5), 879–903. Quine, L. (2003). Workplace bullying, psychological distress, and job satisfaction in junior doctors. Cambridge Quarterly of Healthcare Ethics, 12(1), 91–101. Rayner, C. (1997). The incidence of workplace bullying. Journal of Community & Applied Social Psychology, 7, 199–208. Rayner, C., Hoel, H., & Cooper, C. L. (2002). Workplace bullying: What we know, who is to blame, and what can we do? London: Taylor & Francis. Rodríguez-Muñoz, A., Baillien, E., De Witte, H., Moreno-Jiménez, B., & Pastor, J. C. (2009). Cross-lagged relationships between workplace bullying, job satisfaction and engagement: Two longitudinal studies. Work and Stress, 23(3), 225–243. Roseman, I. J. (1991). Appraisal determinants of discrete emotions. Cognition and Emotion, 5(3), 161–200. Sheehan, M., McCarthy, P. Barker, M., & Henderson, M. (2001, June–July). A model for assessing the impact and costs of workplace bullying. Paper presented at the Standing Conference on Organizational Symbolism (SCOS), Trinity College, Dublin, Ireland. Spector, P. E. (1997). Job satisfaction: Application, assessment, cause, and consequences. London: Sage. Tracy, S. J., Lutgen-Sandvik, P., & Alberts, J. K. (2006). Nightmares, demons and slaves: Exploring the painful metaphors of workplace bullying. Management Communication Quarterly, 20, 148–185. Tugade, M. M., Fredrickson, B. L., & Barrett, L. F. (2004). Psychological resilience and positive emotional granularity: Examining the benefits of positive emotions on emotion regulation and health. Journal of Personality, 72(6), 1161–1190. Van Eck, M., Nicolson, N. A., & Berkhof, J. (1998). Effects of stressful daily events on mood states: Relationship to global perceived stress. Journal of Personality and Social Psychology, 75(6), 1572–1585. Van Veldhoven, M., & Meijman, T. F. (1994). Het meten van psychosocialea rbeidsbelasting met een vragenlijst: de vragenlijst beleving en beoordeling van de arbeid (VBBA) [The measurement of psychosocial job demands with a questionnaire: The questionnaire on the experience and evaluation of work (QEEW)]. Amsterdam: Nederlands Instituut voor Arbeidsomstandigheden. Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54(6), 1063–1070. Weiss, H. M., & Beal, D. J. (2005). Reflections on affective events theory. In N. M. Ashkanasy, W. J. Zerbe, & C. E. J. Härtel (Eds.), Research on emotion in organization: The effect of affect in organizational settings (pp. 1–22). Oxford, United Kingdom: Elsevier/JAI.

172

Glasø and Notelaers

Weiss, H. M., & Cropanzano, R. (1996). Affective events theory: A theoretical discussion of the structure, causes, and consequences of affective experiences at work. Research in Organizational Behavior, 18, 1–74. Williams, K. D. (1997). Social ostracism. In R. M. Kowalski (Ed.), Aversive interpersonal behaviors (pp. 133–170). New York: Plenum Press. Correspondence regarding this article should be directed to Lars Glasø, PhD, BI Norwegian Business School, Nydalsveien 37, 0484 Oslo, Norway. E-mail: [email protected]