Research Presentation

AGGRESSIVE BEHAVIOR

Volume 32, pages 376–384 (2006)

A Developmental Perspective on Bullying Debra J. Pepler1,4�, Wendy M. Craig2, Jennifer A. Connolly1, Amy Yuile1, Loren McMaster3, and Depeng Jiang1

1LaMarsh Centre for Research on Violence and Conflict Resolution, York University, Toronto, Ontario, Canada 2 Department of Psychology, Queen’s University, Kingston, Ontario, Canada

3BC Children’s Hospital, Psychology Department, Vancouver, British Columbia, Canada 4Community Health Systems Resource Group, Hospital for Sick Children, Toronto, Ontario, Canada

: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :

In this paper, we examined the forms and relationship contexts of bullying in adolescence. Using cross-sectional data, we assessed grade and sex differences in self-reports of bullying and sexually harassing peers, as well as reports of dating aggression from1896 students from early to late adolescence. Reports of bullying others were highest around the school transition, with lowest levels at the end of high school. Boys reported more bullying and sexual harassment than girls. Sexual harassment of same- and opposite-sex peers increased over the early adolescent years and leveled off in later high-school years. There were no sex differences in the prevalence of indirect or physical aggression with a dating partner. Adolescents who bullied were at increased risk for the other forms of relationship aggression. These data highlight bullying as a relationship problem and point to the need for prevention programs to curtail the use of power and aggression in adolescent relationships. Aggr. Behav. 32:376–384, 2006. r 2006 Wiley-Liss, Inc.

: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :

INTRODUCTION

Bullying has been defined as negative actions— physical or verbal—that have hostile intent, are repeated over time, and involve a power differential between the bully and the victim [Olweus, 1993]. Through the past 15 years of our research program on bullying, we have come to understand bullying as a relationship problem—because it is a form of aggression that unfolds in the context of a relation- ship in which one child asserts interpersonal power through aggression. The power that bullies hold over others can arise from their individual char- acteristics, such as superior size, strength, or age [Olweus, 1993]; and from knowledge of others’ vulnerabilities [Sutton et al., 1999]. The power in bullying can also arise from a position in a social group, either in terms of generally high social status [Olweus, 1993] or by membership in a group of peers who support bullying [Salmivalli et al., 1997]. In this paper, we examine bullying from a

developmental perspective with a cross-sectional study from early through late adolescence. The two main objectives were to compare grade and sex differences in the prevalence of bullying, sexual harassment, and dating aggression; and to compare adolescents who do and do not report bullying on reports of perpetrating sexual harassment and dating aggression. We contend that the combined use of power and aggression found in playground

bullying is a key underlying component of sexual harassment, dating aggression, workplace harass- ment, marital aggression, and elder abuse [Pepler et al., 1997]. Thus, there is heterotypic continuity in bullying behavior: with development, it transforms into other aggressive behaviors characterized by power and aggression within the context of a relationship. We believe that bullying merits atten- tion because it may underlie many of the problems related to interpersonal violence in our society. From a developmental perspective, we are con- cerned that the lessons learned in bullying within peer relationships generalize to other developmen- tally significant relationships. Few studies examine the developmental pattern

in the prevalence of bullying others beyond early adolescence. Large-scale surveys of the prevalence of bullying reveal a decrease in bullying from childhood through adolescence [Nansel et al., 2001], with evidence of an increase when children

Published online in Wiley InterScience (www.interscience.wiley.

com). DOI: 10.1002/ab.20136

Received 2 February 2005; Accepted 26 April 2005

Grant sponsors: Ontario Mental Health Foundation, The National

Health Research and Development Program, The Social Sciences

and Humanities Research Council of Canada.

� Correspondence to: Debra Pepler, LaMarsh Centre for Research on

Violence and Conflict Resolution, York University, 4700 Keele Street,

Toronto, Ontario, Canada M4R 1P5. E-mail: [email protected]

r 2006 Wiley-Liss, Inc.

make the transition from elementary to high school [Pellegrini and Long, 2002]. In this study, we assessed whether adolescents who bully their peers are more likely to engage in other forms of power and aggression compared to adolescents who do not report bullying. The bio-psycho-social model of development

[Cairns, 1979; Ford and Lerner, 1992; Magnusson, 1988] forms the theoretical foundation for our research. According to this model, the dramatic biological and social changes that occur in adoles- cence impact psychosocial development [Brooks- Gunn et al., 1985]. We expected that pubertal changes in early adolescence are reflected in the emergence of sexual harassment as a form of aggression used to gain power and control. We expected some sex differences in the various forms of bullying. We also expected the changing social relationship contexts from same-sex to mixed-sex peer groups in pre- to early adolescence [Connolly et al., 1999] to be reflected in the emergence of bullying within romantic relationships during adole- scence. In these relationships, power and control may be established through physical aggression, but also through psychological abuse and other forms of behavior that cause distress.

Diversification in Form and Social Relationship context with Age

Consistent with Moffitt’s [1993] notion of hetero- typicality, we hypothesized that the propensity to use various forms of aggression changes with development as a function of increasing capacities and emerging issues. Bjorkqvist et al. [1992] indicate that physical, verbal, and indirect aggression are developmentally sequenced and linked to advances in language and perspective-taking abilities. We propose that the form of bullying also changes as a function of developmental transformations during puberty. During early adolescence, there is an increased awareness of emerging sexuality and sexual identity. Adolescents can readily acquire power over another by identifying vulnerabilities relating to sexuality and, in turn, use these as a means to bully, through sexual harassment. Our research has shown that the perpetration of cross- sex sexual harassment increases through the early adolescent years and is linked to pubertal develop- ment and sex composition of the peer network [Craig et al., 2001; McMaster et al., 2002]. Pellegrini [2001] found an association between bullying and sexual harassment, mediated by dating frequency. We hypothesized, therefore, that sexual harassment

emerges during the early adolescent years and conti- nues through adolescence as a means of asserting power through aggression. We expected that adolescents who bully their peers are more likely to report also perpetrating sexual harassment than those who do not bully. The relationship context of bullying may also

change in concert with biological and social trans- formations during adolescence. A growing interest in romantic relationships accompanies pubertal development. We were interested in the general- ization of power and aggression from bullying within peer relationships to bullying within the context of romantic relationships. We found that students in Grades 6–8 (aged 11–13) who reported bullying others were more likely to be involved in a romantic relationship and more likely to report being verbally and physically aggressive with their romantic partners than students who did not report bullying [Connolly et al., 2000]. The sexual dimen- sion within romantic relationships provides a found- ation from which either male or female partners can exert power and control [Capaldi and Gorman- Smith, 2003]. Adolescents who bully may recognize and target the vulnerabilities in a romantic partner to establish interpersonal power in the relationship to a greater extent than adolescents who do not bully. Thus, we hypothesized that adolescents who report bullying are more likely to report perpetrat- ing aggression towards their romantic partner.

Sex Differences in Bullying

Within the present developmental analysis, we were interested in sex differences in the various forms of power and aggression during adolescence. Girls’ aggressive behavior problems are generally less prevalent than those of boys [Archer, 2004; Moffitt et al., 2001; Offord et al., 2001]. Although the vast majority of studies of bullying use self- report questionnaires, there are measurement con- cerns with form of assessment because girls may be less likely to acknowledge perpetrating this form of aggression than boys. In survey research, girls are less likely to report bullying others compared to boys [Charach et al., 1995; Olweus, 1993; Pepler et al., 2004]. In our observational research, the difference between the rate of boys’ and girls’ bullying was not as great as the survey reports indicated [Pepler et al., 2004]. Although a smaller proportion of girls than boys acknowledge exerting power through bullying, there appears to be a social cost for both girls and boys who bully—they report less closeness and more conflict in relationships

377Development of Bullying

Aggr. Behav. DOI 10.1002/ab

compared with children who do not report bullying others [Pepler et al., 2004]. Given the relationship dynamic in bullying, we are also concerned that girls and boys who use their power aggressively with peers will transfer this dynamic to interactions with romantic partners. There is evidence from our observational research

that girls who bully may be establishing patterns of power and aggression in the context of both same- and opposite-sex relationships. We observed elemen- tary aged girls bullying boys and other girls at approximately equal rates: In our first study, 52% of the victims of girls’ bullying were boys and 48% were girls [Craig and Pepler, 1997]. In our second study, we observed girls bullying boys in 45% and other girls in 55% of the episodes in which girls were the aggressors [Pepler et al., 1998]. In both studies, boys were much more likely to bully other boys than to bully girls, which may relate both to the nature of boys’ interactions, as well as to a social norm against male to female aggression [Archer, 2000]. Research on aggression within adolescent romantic relation- ships reveals few sex differences: girls are as likely to perpetrate dating aggression with their partners as are boys [Connolly et al., 2000]. In a meta-analysis of sex differences in partner aggression, Archer [2000] found that among young, dating couples, females tended to report using more physical aggression than males; however, the effect size was modest. Within a romantic relationship, Archer notes that there is a strong norm of men not hitting women, which may enable the women to engage in physical aggression with little fear of retaliation. We hypothesized, therefore, that boys are more likely to report higher levels of bullying and sexual harassment than girls, but that boys and girls do not differ in reports of aggression within dating relationships.

METHOD

Cross-sectional data for this study were drawn from a study of bullying in late elementary school and high school. There were two independent samples of students: one from elementary school (Grades 6–8) and the other from high school (Grades 9–12). The schools were located in a large Canadian city. These cross-sectional samples were part of a larger study with common measures of the various forms of aggression across the two studies.

Participants

Elementary school sample. This sample com- prised 504 boys and 457 girls enrolled in Grades 6–8

in seven elementary schools (Kindergarten through Grade 8), with an average age of 12.6 years (SD 5 .9 years; range 9–14). There were 114 boys and 126 girls in Grade 6, 169 boys and 130 girls in Grade 7, and 223 boys and 201 girls in Grade 8. Approxi- mately 75% of the students had parental consent to participate in the study. Five schools were coeduca- tional and publicly funded and two schools were single-sex independent schools, one for boys only and one for girls only. The majority of students were from Euro-Canadian backgrounds; however, speci- fic data on ethnicity are not available as this information was not collected for the elementary school sample. Most of the adolescents came from two-parent households (74.6%). Of the remaining adolescents, 4.0% lived with both parents separately in joint custody arrangements, 4.7% lived in blended families (with one biological parent and one step- parent), 14.2% came from single-parent homes, and 2.4% were living in other types of family configura- tions (e.g., with legal guardians). Students from the public schools were from lower- and middle-socio- economic level families, whereas students from the independent schools were from middle- to upper- socioeconomic level families. Specific information about parental level of education was not collected from this sample.

High-school sample. The high-school sample comprised 456 boys and 479 girls enrolled in Grades 9–12 in four high schools, with an average age of 15.7 years (SD 5 1.2 years; range 13–19). There were 142 boys and 108 girls in Grade 9, 117 boys and 129 girls in Grade 10, 132 boys and 105 girls in Grade 11, and 66 boys and 137 girls in Grade 12. In each of the schools, approximately 65% of the students had parental consent to participate in the study. Two high schools were coeducational publicly funded schools and two schools were single-sex independent schools, one for boys only and one for girls only. The majority of students were from Euro-Canadian backgrounds (70.6%), with 5.0% African-Canadian adolescents, 13.3% Asian-Canadian adolescents, and 11.1% from other ethnic backgrounds. Most of the adolescents lived in households with both biological parents (72.6%). Of the remaining ado- lescents, 5.9% lived with both parents separately in joint custody arrangements, 5.5% lived in blended families (with one biological parent and one step- parent), 14.4% came from single-parent homes, and 2.9% were living in other types of family configurations. Eleven percent of the adolescents’ fathers and 14% of their mothers had completed high school, while 73.8% of the fathers and 69.7%

378 Pepler et al.

Aggr. Behav. DOI 10.1002/ab

of the mothers had gone on to complete at least some post-secondary education.

Measures

Bullying. A shortened version of The Safe School Questionnaire, adapted from Olweus [1989], was administered to obtain self-report of bullying behavior in school. Two items were used to assess involvement in bullying: ‘‘How often have you taken part in bullying others in the past 2 months?’’ scored with a five-point scale (0 5 not at all, 1 5 once or twice, 2 5 now and then, 3 5 about once a week, 4 5 several time a week); and ‘‘How often have you taken part in bullying others in the last 5 days?’’ scored with a five-point scale (0 5 not at all, 1 5 once, 2 5 twice, 3 5 three or four times, 4 5 five or more times). A sum score of the two items was used as a measure of bullying behavior. Prior to administration, a class discussion was conducted to define bullying and a definition of bullying, adapted from Olweus [1989], was provided. This definition clarified that bullying was repeated aggression and that the victimized child had difficulty defending him/herself. The definition did not specify sexual forms of bullying, such as sexual harassment. The internal reliability for the bullying questions was .84. The use of these two items to assess bullying has

limitations. Without a listing of the behaviors that might be perpetrated in bullying, such as hitting (physical), name-calling (verbal), and excluding (social), there may be an under-reporting of involvement in bullying. Given that bullying is more consistent with the traditional male stereotype, boys may be more willing to report bullying than girls. Our observations of bullying on the school play- ground suggest that the discrepancy between girls’ and boys’ involvement in bullying may not be as great as self-reports imply [Pepler et al., 2004]. Although self-reports of bullying using this form of question have been used extensively, these measurement limitations should be kept in mind in the consideration of sex differences in bullying.

Sexual harassment. Sexual harassment was measured using a modified version of the AAUW Sexual Harassment Survey [AAUW, 1993], which asks students to report on how often they perpe- trated a variety of sexual harassment behaviors. The questionnaire instructions make explicit that the students are to report on only unwanted sexual behaviors. The following modifications were made to the original survey: (1) students reported only on harassment involving peers, not staff; (2) students reported on harassment occurring in the last

2 months, not their entire school lives; and (3) the response rating scale was expanded to five points, with the anchors ranging from 0 5 never through 4 5 daily. [For more details, see McMaster et al., 2002]. Five items were included: making sexual comments, jokes, movements, or looks; brushing up against someone in a sexual way on purpose; spreading sexual rumors about someone; calling someone a ‘‘fag,’’ ‘‘dyke,’’ ‘‘lezzie,’’ or ‘‘queer’’; flashing or mooning someone. Because of our interest in the distinction between same-sex and opposite-sex harassment, the students were asked, for each item, how often the behavior was directed to a same-sex and to an opposite-sex peer. Because the item distributions were strongly positively skewed, items were dichotomized into 0 5 ‘‘never’’ and 1 5 ‘‘ever.’’ Internal reliabilities for dichoto- mous coding were .78 and .86 for same- and opposite-sex perpetration, respectively.

Dating aggression. To assess dating aggres- sion, adolescents responded to seven items with reference to behavior with either a current or recent (past 3 months) romantic partner. These items were the same as those for aggression with same- and opposite-sex peers. Three items from the Conflict Tactics Scale [Straus, 1979] assessed physical aggres- sion (‘‘slapped or kicked’’; ‘‘choked, punched, or beaten’’; ‘‘threatened with a knife’’). Three items from the Relational Aggression Scale [Crick and Grotp- eter, 1995] assessed indirect aggression (‘‘spread rumors or lies about him/her’’, ‘‘when mad, kept him/her out of the group’’, ‘‘ignored him/her when mad’’). Using a five-point scale, ranging from ‘‘never happened’’ to ‘‘happened more than nine times’’ adolescents reported on perpetrating these behaviors with a boy/girlfriend. a coefficients were .94 and .83 for physical and indirect aggression, respectively.

Pubertal development. Pubertal status was measured in grades 6–8 using the Pubertal Develop- ment Scale [Petersen et al., 1988]. On sex-specific versions of the form, girls and boys rated the develop- ment of their secondary sex characteristics (ranging from not yet started to completed) including pubic hair, growth spurt, skin changes, facial hair, voice change, breast development, and menarchial status. a coefficients were .79 for boys and .83 for girls.

Procedure

Trained research staff administered the question- naires during regularly scheduled class periods. All participating students had obtained signed parental consent and assented themselves to participate. Because of the sensitive nature of the research, we

379Development of Bullying

Aggr. Behav. DOI 10.1002/ab

included a question at the end of the package invit- ing students to indicate if they wanted to talk to a member of the research team about issues raised in the questionnaires. We developed a protocol to address the issues that adolescents raised in discussions with senior research staff [cf. Yuile et al., in press].

RESULTS

Grade and Sex Differences in the Form and Relationship context of Bullying and Related Behaviors

To examine the developmental patterns of bully- ing and aggression, we assessed grade and sex differences on summary scores on the bullying, sexual harassment, and dating aggression items. We used grade, rather than age, as an indicator of developmental stage to reflect the influence and importance of school groupings and peer group contexts in the social behaviors of adolescents. To assess grade and sex differences in the patterns

of aggression, we conducted two MANOVAs: one for the scale scores for bullying and same- and opposite-sex sexual harassment; the other for

indirect and physical aggression with dating part- ners. In the first analysis, there was a multivariate effect of grade, F(18, 4,614) 5 7.01, Po.001 and of sex, F(3, 1,631) 5 45.71, Po.001, with no grade by sex interaction. The grade and sex means for all forms of aggression are presented in Table I. Reports of bullying were lower in the elementary grades (6–8) than in the high-school grades (9–12), with a peak at the school transition point (grade 9), followed by lower reports of bullying for students in grades 10 and 12, F(6, 1,633) 5 4.57, Po.001. There were also grade differences in reports of sexual harassment towards same-sex peers, F(6, 1,633) 5 13.99, Po.001, and opposite-sex peers, F(6, 1,633) 5 11.30, Po.001. Consistent with a deve- lopmental hypothesis, reports of sexually harassing same-sex and opposite-sex peers were lower in the elementary grades (6–8) than in high-school grades (9–11). Reports of same-sex harassment were some- what lower among grade 12 students compared to other high-school students and not different from those reported by students in grades 7 and 8. Reports of opposite-sex harassment were highest among grade 9 students and significantly higher than reports by students in grades 6–8. Reports of

TABLE I. Mean Aggression Scores (Scale 1–5) (and Confidence Intervals) by Grade and Sex

Bullying SS sexual harassment OS sexual harassment Social dating aggression Physical dating aggression

Grade 6

Boys 1.25 [1.13–1.43] 1.11 [1.03–1.18] 1.05 [.96–1.14] 1.02 [.91–1.13] 1.03 [.96–1.09]

Girls 1.16 [1.02–1.30] 1.16 [.94–1.09] 1.15 [.94–1.13] 1.10 [.92–1.14] 1.06 [.94–1.08]

Total 1.21a [1.12–1.32] 1.06a [1.01–1.12] 1.04a [.98–1.11] 1.02a [.95–1.10] 1.02 [.97–1.06]

Grade 7

Boys 1.43 [1.33–1.58] 1.16 [1.10–1.23] 1.15 [1.07–1.23] 1.10 [1.02–1.18] 1.06 [1.02–1.11]

Girls 1.17 [1.03–1.32] 1.02 [.95–1.10] 1.06 [.97–1.16] 1.02 [.93–1.12] 1.01 [.95–1.06]

Total 1.31 [1.22–1.41] 1.11a [1.04–1.15] 1.11a [1.05–1.17] 1.07a [1.00–1.13] 1.04 [1.00–1.07]

Grade 8

Boys 1.57 [1.47–1.68] 1.21 [1.15–1.27] 1.25 [1.18–1.31] 1.09 [1.02–1.15] 1.07 [1.04–1.11]

Girls 1.26 [1.17–1.41] 1.04 [.97–1.10] 1.10 [1.02–1.17] 1.09 [1.01–1.17] 1.03 [.98–1.07]

Total 1.42b 1.35–1.51] 1.14a [1.08–1.17] 1.19b [1.12–1.22] 1.09a [1.04–1.14] 1.05 [1.02–1.08]

Grade 9

Boys 1.63 [1.52–1.76] 1.41 [1.34–1.47] 1.40 [1.33–1.48] 1.21 [1.13–1.29] 1.02 [.97–1.07]

Girls 1.39 [1.22–1.50] 1.18 [1.10–1.25] 1.31 [1.18–1.36] 1.38 [1.27–1.50] 1.08 [1.02–1.15]

Total 1.53b [1.41–1.59] 1.31b [1.24–1.34] 1.36c [1.30–1.42] 1.27c [1.21–1.33] 1.04 [1.01–1.09]

Grade 10

Boys 1.48 [1.35–1.62] 1.44 [1.37–1.51] 1.36 [1.28–1.45] 1.20 [1.08–1.26] 1.07 [1.01–1.13]

Girls 1.13 [1.00–1.25] 1.16 [1.09–1.22] 1.21 [1.12–1.29] 1.21 [1.11–1.31] 1.03 [.98–1.09]

Total 1.30a [1.21–1.40] 1.29b [1.25–1.35] 1.28c [1.23–.34] 1.20b [1.12–1.26] 1.05 [1.01–1.09]

Grade 11

Boys 1.61 [1.49–1.74] 1.38 [1.31–1.45] 1.36 [1.28–1.44] 1.20 [1.12–1.29] 1.05 [1.0–1.10]

Girls 1.15 [1.00–1.29] 1.09 [1.02–1.17] 1.17 [1.08–1.26] 1.30 [1.20–1.39] 1.05 [.99–1.10]

Total 1.41b [1.29–1.48] 1.25b [1.19–1.29] 1.27c [1.21–1.33] 1.24c [1.18–1.30] 1.05 [1.01–1.09]

Grade 12

Boys 1.33 [1.16–1.51] 1.35 [1.25–1.44] 1.32 [1.21–1.44] 1.36 [1.25–1.47] 1.06 [1.00–1.13]

Girls 1.11 [.99–1.23] 1.14 [1.07–1.21] 1.18 [1.11–1.27] 1.20 [1.12–1.28] 1.03 [.98–1.08]

Total 1.18a [1.11–1.33] 1.21c [1.19–1.30] 1.23b [1.16–1.30] 1.26c [1.20–1.32] 1.04 [1.00–1.09]

Frequencies for the total score for form of aggression with different letter subscripts are significantly different between grades (Po.05).

380 Pepler et al.

Aggr. Behav. DOI 10.1002/ab

opposite-sex harassment were somewhat lower during the latter high-school grades. Grade 10 and 11 students’ reports were not different from those of grade 8 students and grade 12 students’ reports were similar to those of students in grades 7 and 8. The multivariate effect of sex revealed that boys

were more likely to report perpetrating aggression than girls. As hypothesized, boys reported signi- ficantly higher levels of bullying, F(1, 1,633) 5 61.13, Po.001, sexual harassment toward same-sex peers, F(1, 1,633) 5 103.75, Po.001, as well as toward opposite-sex peers, F(1, 1,633) 5 27.69, Po.001. The second MANOVA assessed indirect and

physical aggression with dating partners. We exam- ined dating aggression separately from bullying and sexual harassment because not all of the adolescents had a current or recent dating partner. Table I presents the mean scores of boys and girls in the elementary and high-school samples. A multivariate effect was found for grade, F(12, 1,964) 5 6.71, Po.001, but not for sex, with no grade by sex interaction. Overall, levels of indirect aggression towards dating partners were significantly lower in the elementary grades (6–8) than in the high-school grades (9–12), F(6, 983) 5 10.64, Po.001. Reports of indirect aggression in dating relationships were somewhat low in Grade 10; however, the overall level of aggression in Grade 10 did not differ from the level reported by students in Grades 7 and 8. The main effect for grade did not hold up in the univariate analysis of physical aggression within dating relationships. The prevalence of physical aggression toward a dating partner was generally low across grades, ranging from 2% to 24%. An examination of the confidence intervals in Table I indicates that for indirect and physical dating aggression, girls’ scores were similar to those of boys. The exception to the pattern of no sex differences arose in Grade 9, when the girls’ mean for indirect dating aggression was higher than that for boys, and in Grade 12 when the boys’ mean for indirect dating aggression was higher than that for girls. We have postulated that sexual harassment and

dating aggression are forms of bullying that emerge in concert with puberty when early adolescents experience both physical and psychosocial sexual development. Pubertal development was assessed for students in grades 6–8. Although there were no significant correlations between pubertal develop- ment and same-sex sexual harassment, several correlations between puberty and opposite-sex sexual harassment were significant. For girls, the correlations for grades 6, 7, and 8 were significant

(.31, .24, and .24, respectively). For boys, only the grade 7 correlation was significant (.21). For dating aggression, the only significant correlations were again for boys in grade 7: pubertal development correlated significantly with grade 7 boys’ reports of both physical aggression (.311) and indirect aggression (.316) toward a romantic partner.

Risk of Engaging in Aggression Among Students who Bully

The second objective was to assess the likelihood of engaging in behaviors that represent a develop- mental diversification of the form and relationship context of bullying: sexual harassment and dating aggression. We first examined the correlations between students’ reports of bullying and engaging in these other forms of power and aggression. Elementary and high-school boys’ and girls’ reports of bullying were significantly correlated with their reports of sexual harassment and dating aggres- sion, ranging from .16 to .44. Therefore, there is consistency in students’ reports of using power and aggression: high scores on bullying were associated with high scores on sexual harassment and dating aggression. To assess the risk of engaging in other forms of relationship aggression for adolescents who bully, we ran four w2 analyses, separately for boys and girls in elementary and high school, which are presented in Table II. The probability of involvement in sexual harassment and dating aggression was compared between youth who reported bullying others at least once in the past 2 months and those who did not report bullying their peers.

Sexual harassment. There were significant associations between the level of bullying and boys’ and girls’ reports of same- and opposite-sex sexual harassment in both elementary and high school. Sexual harassment was more prevalent among students who bullied than those who did not report bullying others. The patterns were consistent for both elementary and high school: both boys and girls who reported bullying others were more likely to report sexually harassing same-sex peers and opposite-sex peers compared to boys and girls who did not bully.

Dating aggression. There were also significant associations between bullying and indirect and physical dating aggression. In general, the elemen- tary and high-school boys and girls who reported bullying their peers were more likely to report perpetrating both forms of dating aggression than those who did not bully. The exception to this

381Development of Bullying

Aggr. Behav. DOI 10.1002/ab

pattern was for high-school girls’ indirect dating aggression: the proportion of non-bullying girls (40%) was similar to the proportion of bullying girls (43%) who reported being indirectly aggressive with their dating partners.

DISCUSSION

In our research program on bullying, we have come to understand bullying as a relationship problem, because it unfolds in the context of rela- tionships. For children who bully, we are concerned that the interactional style of using power and aggression will generalize to other forms of relation- ship aggression. In this paper, we have examined the age and sex differences in reports of bullying and the changing form and relationship context of the use of power and aggression in a cross-sectional study of adolescents. Consistent with developmental expecta- tions, the data highlight the increase and decrease in reports of bullying across the adolescent years and the emergence of sexual harassment and dating aggression. The emergence of these forms of beha- vior coincides with pubertal development, when issues of sexuality are salient. Consistent with this developmental pattern, we found that those children who are more advanced in pubertal development are more likely to use some of these emerging forms of bullying. Opposite-sex sexual harassment was asso- ciated with pubertal development for girls in grades 6–8; for boys, opposite-sex sexual harassment was only significant for boys in grade 7. The association between dating aggression and pubertal develop- ment was also found only for boys in grade 7. The present data suggest an association between advan- ced pubertal timing and norm-breaking behaviors. For girls, aggression is inconsistent with social

norms, but the well-documented sex difference in aggression appears to hold only for aggression to a same-sex and not to an opposite-sex person [Archer, 2004]. Therefore, girls who are advanced in pubertal development may risk the general norm violations for aggression to engage, albeit negatively, with boys during early adolescence when interest in the opposite sex is increasing. We observed a similar form of negative engagement on the school play- ground by younger girls [Pepler et al., 2004]. For boys, advanced pubertal development relative to peers was also associated with norm violations: they were more likely to be aggressive toward a girl through sexual harassment or physical and social dating aggression. This pattern was only seen with boys in grade seven, however, suggesting that the boys who are most advanced in pubertal develop- ment in this early stage may be at particular risk for relationship aggression. There is a need for more research to explore these associations. The present data suggest that the timing of pubertal develop- ment may be important in association with relation- ship patterns that bode poorly for both boys’ and girls’ healthy opposite-sex relationships through adolescence and into adulthood. The grade differences found in this cross-sectional

analysis of bullying and sexual harassment are only somewhat consistent with those found for other forms of aggression. There was a higher level of bullying reported among the high school than among the elementary school students. This pattern is inconsistent with the findings of Nansel et al. [2001], who reported cross-sectional data indicating higher levels in bullying reported by children than by adolescents. An inspection of our data reveals that the prevalence of bullying was highest for boys in Grade 8 and girls in Grade 9, when almost half of the students reported bullying others at least

TABLE II. The Association between Bullying and Other Forms of Aggression for Boys and Girls in Elementary and High School

Elementary school High school

Non-bully Bully w2 Non-bully Bully w2

Boys

SS sex harass 55/299 (18%) 90/197 (46%) 42.7 ���

24/119 (20%) 50/100 (50%) 21.6 ���

OS sex harass 27/301 (9%) 69/197 (35%) 51.9 ���

18/121 (15%) 43/100 (43%) 21.7 ���

Soc date agg 12/153 (8%) 29/121 (24%) 13.8 ���

3/62 (5%) 17/68 (25%) 10.1 ��

Phys date agg 6/156 (4%) 12/122 (10%) 4.1 �

2/63 (3%) 9/68 (13%) 4.3 �

Girls

SS sex harass 7/252 (3%) 17/78 (22%) 31.9 ���

97/369 (26%) 51/95 (54%) 26.1 ���

OS sex harass 26/254 (10%) 33/78 (42%) 42.0 ���

137/374 (37%) 66/101 (66%) 26.8 ���

Soc date agg 11/150 (7%) 13/60 (22%) 8.7 ��

77/193 (40%) 24/56 (43%) n.s.

Phys date agg 1/145 (1%) 5/57 (9%) 9.3 ��

23/195 (12%) 13/56 (23%) 4.6 �

Note: ***Po.001, **Po.01, *Po.05.

382 Pepler et al.

Aggr. Behav. DOI 10.1002/ab

once in the past 2 months. In a longitudinal analysis, Pellegrini and Long [2002] found that bullying increased when children made the transition from elementary to high school. Although cross-sectional, the data in the present study indicate that the years on either side of a school transition may be a period of risk for bullying. For the boys in our sample, the peak in reports of bullying appeared at the time when they were the oldest in elementary school, whereas for the girls, the peak appeared as they were adjusting to the new social context of high school with new relationships to be established. Our data are consistent with those of Nansel et al. [2001] indicating that reports of bullying were low at the end of high school. The pattern of generally low levels of aggression for older adolescents is con- sistent with adolescents’ increasing capacity for empathy and less tolerance of those who are mean and hurtful [Galambos et al., 2003]. Consistent with a bio-psycho-social perspective of

development, sexual harassment emerged as a form of aggression used to gain power and control. The means for both same-sex and opposite-sex sexual harassment increased in a step-wise fashion across the elementary school years to a peak in reporting by both boys and girls in who were Grade 10, after which the means stayed relatively constant. For boys in Grade 10, the prevalence of same-sex sexual harassment (65%) was somewhat higher than that for opposite-sex harassment (56%), whereas for the Grade 10 girls, the prevalence of opposite-sex sexual harassment (43%) was somewhat higher than that for same-sex harassment (36%). There was a developmental pattern in reports of

dating aggression, particularly for indirect aggression, which was more frequently reported by high school than by elementary school students. Björkqvist et al. [1992] highlighted this general developmental progression in the forms of aggression. The present research indicates that the more subtle and sophis- ticated forms of indirect aggression toward a romantic partner, such as ignoring and exclusion, are more prevalent among older adolescents. The grade difference was not significant for physical aggression with a romantic partner. Fewer than 10% of boys and girls in elementary school reported physical aggression with a dating partner. Among high-school students, the rates were somewhat higher, but generally below 15%. The sex differences in bullying and sexual harass-

ment were consistent with those found with other forms of aggression, with more boys acknowledging using these forms of aggression than girls [Archer, 2004; Moffitt et al., 2001]. The exception to this

trend was in dating aggression: there were no sex differences in the prevalence of indirect and physical forms of dating aggression. The context of romantic relationships with high levels of intimacy and opportunity for conflict appears to engage a relatively equal number of boys and girls in aggressive interactions. Indirect aggression is re- ported more frequently than physical aggression in romantic relationships, with over a quarter of high- school boys and girls reporting indirect aggression. The equal rates of boys and girls reporting dating aggression raise concerns for the interactional patterns being established. Dating aggression in these young intimate relationships may foretell longer-term relationship problems and the potential for acceleration into violence. The concern for bullying as a form of behavior in

which children learn to use aggression to establish power in relationships is highlighted in the present study. The cross-sectional data, from an elementary and a high-school cohort, are however limited, in that they cannot be used to test whether there is a developmental progression from bullying to sexual harassment and dating aggression, which requires a prospective study. The present data reveal that the youth who report bullying their peers are also more likely to provide concurrent reports of engaging in developmentally salient forms of aggression—sexual harassment and dating violence. In elementary and high school, bullying is not an exceptional problem: many adolescents engage in bullying their peers occasionally; a small group of students bully at a more frequent rate. The latter group of adolescents, who frequently bully others, may be at high risk for transferring these interactional patterns to other forms of power and aggression, such as workplace harassment, domestic violence, and child abuse. The developmental pathway from bullying to other forms of relationship aggression in adulthood is a critical area for future research. This study of bullying across the adolescent years

draws attention to adolescents who are involved in using aggression to establish power and control over others. Although limited by the self-report and cross-sectional nature of the data, the present study highlights the importance of identifying and inter- vening with students involved in bullying: those adolescents who bully are more likely to sexually harass their same- and opposite-sex peers and are more likely to be physically aggressive with their dating partners. These boys and girls need to be identified as early as possible to stop the reinforcing dynamics that emerge in bullying and to prevent the further consolidation of their aggressive inter-

383Development of Bullying

Aggr. Behav. DOI 10.1002/ab

actional patterns. When bullying is understood as a relationship problem, a clear objective for inter- vention can be identified: to enhance relationship capacity and promote healthy relationships in the present in order to lay the foundation for healthy relationships throughout the lifespan.

ACKNOWLEDGMENTS

The authors appreciate the children, adolescents, schools, and parents who participated in the research. They are indebted to Arland O’Hara, the Research Coordinator, and to the many graduate students and undergraduate research assistants for their contributions and dedication to the research program.

REFERENCES

AAUW. 1993. ‘‘Hostile hallways. The AAUW survey on sexual

harassment in America’s schools,’’ Washington, DC: American

Association of University Women Educational Foundation.

Archer J. 2000. Sex differences in aggression between heterosexual

partners: A meta-analytic review. Psychological Bulletin

126:651–680.

Archer J. 2004. Sex differences in aggression in real-world settings: A

meta-analytic review. Rev Gen Psychol 8:291–322.

Björkqvist K, Osterman K, Kaukiainen A. 1992. The development of

direct and indirect aggressive strategies in males and females. In:

Björkqvist K, Niemela P (eds): ‘‘Of Mice and Women: Aspects of

Female Aggression,’’ San Diego, CA: Academic Press, pp 51–64.

Brooks-Gunn J, Peterson AC, Eichorn D. 1985. The study of

maturational timing effects in adolescence. J Youth Adolescence

14:149–161.

Cairns RB. 1979. ‘‘The Analysis of Social Interactions: Methods,

Issues, and Illustrations,’’ Hillsdale, NJ: Erlbaum.

Capaldi DM, Gorman-Smith D. 2003. Physical and psychological

aggression in male/female adolescent and young adult couples.

In: Florsheim P (ed): ‘‘Adolescent Romantic Relations and

Sexual Behavior: Theory, Research and Practical Implications,’’

Mahwah, NJ: Erlbaum, pp 244–278.

Charach A, Pepler D, Ziegler S. 1995. Bullying at school:

A Canadian perspective. Educ Can 35:12–18.

Connolly J, Goldberg A, Craig W, Pepler D. 1999. Development and

significance of cross-sex activities in early adolescence. J Youth

Adolescence 28:481–494.

Connolly J, Pepler D, Craig W, Taradash A. 2000. Dating

experiences of bullies in early adolescence. Child Maltreatment 5:

299–310.

Craig WM, Pepler DJ. 1997. Observations of bullying and

victimization in the schoolyard. Can J Sch Psychol 13:41–60.

Craig W, Pepler D, Connolly J. 2001. Developmental context of peer

harassment in early adolescence: The role of puberty and the peer

group. In: Juvonen J, Graham S (eds): ‘‘Peer Harassment in

School: The Plight of the Vulnerable and Victimized,’’ New

York: Guilford, pp 242–261.

Crick N, Grotpeter J. 1995. Relational aggression, gender, and

social–psychological adjustment. Child Dev 66:710–722.

Ford D, Lerner RM. 1992. ‘‘Developmental Systems Theory:

An Integrative Approach,’’ Newbury Park: Sage Publications.

Galambos NL, Barker EV, Tilton-Weaver LC. 2003. Canadian

adolescents’ implicit theory of immaturity: What does childish

mean?. In: Arnett JJ, Galambos NL (eds): ‘‘Exploring Cultural

Conceptions of the Transition to Adulthood. New Directions for

Child and Adolescent Development,’’ vol. 100, pp 77–90.

Magnusson D. 1988. ‘‘Individual Development From an Inter-

actional Perspective: A Longitudinal Study,’’ Hillsdale, NJ:

Lawrence Erlbaum Associates.

McMaster L, Connolly J, Pepler D, Craig W. 2002. Peer-to-peer

sexual harassment among early adolescents: A developmental

perspective. Dev Psychopathol 14:91–105.

Moffitt TE. 1993. Adolescence-limited and life-course-persistent

antisocial behavior: A developmental taxonomy. Psychology 100:

674–701.

Moffitt TE, Caspi A, Rutter M, Silva PA. 2001. ‘‘Sex Differences in

Antisocial Behavior,’’ Cambridge, UK: Cambridge University

Press.

Nansel TR, Overpeck M, Pilla RS, Ruan WJ, Simons-Morton B,

Scheidt P. 2001. Bullying behaviors among US youth: Prevalence

and association with psychosocial adjustment. J Am Med Assoc

285:2094–2100.

Offord DR, Lipman EL, Duku EK. 2001. Epidemiology of problem

up to age 12 years. In: Loeber R, Farrington D (eds): ‘‘Child

Delinquents: Development, Intervention, and Service Needs,’’

Thousand Oaks: Sage Publications; pp 95–116.

Olweus D. 1989. Questionnaire for students (Junior and Senior

versions). Unpublished manuscript.

Olweus D. 1993. ‘‘Bullying at School: What We Know and What We

Can Do,’’ Cambridge, MA: Blackwell Publishers.

Pellegrini AD. 2001 A longitudinal study of heterosexual relation-

ships, aggression, and sexual harassment during the transition

from primary school through middle school. J Appl Dev Psychol

22:119–133.

Pellegrini AD, Long JD. 2002. A longitudinal study of bullying,

dominance, and victimization, during the transition from primary

through secondary school. Br J Dev Psychol 20:259–280.

Pepler D, Craig W, Connolly J. 1997. ‘‘Bullying and Victimization:

The Problems and Solutions for School-aged Children,’’ Ottawa,

Canada: National Crime Prevention Council.

Pepler DJ, Craig WM, O’Connell P, Atlas R. 1998. Observations

of bullying on the school playground. Paper presented at the

XVth Biennial Meeting of International Society for the study

of behavioural Development. Berne, Switzerland.

Pepler D, Craig W, Yuile A, Connolly J. 2004. Girls who bully:

A developmental and relational perspective. In: Putallaz M,

Bierman KL (eds): ‘‘Aggression, Antisocial Behavior, and

Violence Among Girls: A Developmental Perspective,’’ New

York: Guildford, pp 90–109.

Petersen A, Crockett L, Richards M, Boxer A. 1998. A self-report

measure of pubertal status: Reliability, validity, and initial

norms. J Youth and Adolescence 17:117–133. [Scott, BF699D46]

Salmivalli C, Huttunen A, Lagerspetz KMJ. 1997. Peer networks and

bullying in schools. Scand J Psychol 38:305–312.

Straus MA. 1979. Measuring intra-family conflict and violence: The

Conflict Tactics (CT) Scale. J Marriage Family 41:75–88.

Sutton J, Smith PK, Swettenham J. 1999. Social cognition and

bullying: Social inadequacy or skilled manipulation? Br J Dev

Psychol 17:435–450.

Yuile A, Pepler D, Craig W, Connolly J. 2006. The ethics of peeking

behind the fence: Issues related to studying children’s aggression

and victimization. In: Leadbeater B, Nanister E, Benoit C,

Jansson M, Marshall A, Rieckem T (eds). ‘‘Ethics of Research

with High-Risk Children,’’ Toronto: University of Toronto Press,

p 70–89.

384 Pepler et al.

Aggr. Behav. DOI 10.1002/ab