Introduction

Bullying

Bullying is rampant in Delhi/NCR schools with 96 per cent participants in a survey claiming that this kind of abuse is of a significant concern in educational establishments. A recent study conducted by Fortis National Mental Health programme found that “instances of bullying continue to threaten the physical and psychological safety of school-going children year after year”. The survey confirmed that bullying had “a traumatic effect on children” with 61 per cent respondents reported witnessing bullying in classrooms and 75 per cent in school corridors. Also 55 per cent school teachers feel that bullying makes children feel unsafe in schools (The Hindu dated 4 September, 2014).

A video of a student being bullied by a bunch of his schoolmates in a Delhi school has shocked parents who have demanded strict action against the boys. In the video, that has gone viral on mobile, the two class 6 students are seen hurling abuses at another boy, apparently from Class 7, while two others encourage them. The video was shot on Monday by a student on his cellphone, despite the gadget being banned for students within the school premises (ndtv.com dated March, 2014).

The above news reports share one thing in common and that is the sad consequences of bullying behavior prevalent in schools. Bullying is a problem faced worldwide, affecting about 1 in 5 school-aged children (Glew, Rivara, & Feudtner, 2000). Surveys have shown that the proportion of school-aged children who report being bullied is remarkably consistent across countries: Australia (17 %), England (19 %), Japan (15 %), Norway (14 %), Spain (17 %) and the USA (16 %). About 20 % of children report being a bully themselves (Nansel et al., 2001). The prevalence of bullying appears to peak at age 7 (grade 2) and at ages 10–12 (grades 6–8).

The study cited above happens to be one of the few researches carried out in the Indian context. The limited research on bullying seems to reflect that the schools are in the grip of this problem, though the challenge has not been completely comprehended or confronted. More often than not incidents of school violence and aggression capture the attention of all concerned, but for various compulsions, the matter is not dealt with the way it should be.

Although bullying may be traced back to 1800s, research on the subject did not begin until the 1970s. It was with the work of Olweus (1993) that the term “bullying” came to be understood afresh. According to him, bullying includes: (1) A power imbalance in which the child doing the bullying has more power because of age, size, support of peer group or higher status; (2) it is carried out with the intention of harming the targeted child; and (3) is an activity in which a particular child is singled out repeatedly. According to Kenneth Rigby (2002) and other developmental psychologists (Campbell, 2005; Olweus, 2001; Whitney & Smith, 1993), bullying involves a desire to hurt, a harmful action, a power imbalance (typically) repetition, an unjust use of power and an evident enjoyment by the aggressor and generally a sense of being oppressed on the part of the victim.

Bullying occurs in various forms (physical, verbal and social/relational), some of which are vivid and others of which are very subtle, and cause emotional and psychological harm. Broadly, it occurs in two forms: direct bullying and indirect bullying. In cases of direct bullying (which involves use of physical and verbal aggression), the child, and often others, knows the identity of person(s) doing the bullying. This makes bullying easily recognizable, as the behavior is readily observable and the impact is immediate. In contrast, indirect bullying includes covert, harmful behaviors directed toward another child (Olweus, 1993). These might involve social or relational aggressive acts or threats of excluding victims from peer groups or social interactions (Cornell, 2006). Although boys and girls participate in both forms of bullying behavior, boys are more likely to be involved in direct bullying and girls in relational bullying.

With respect to the characteristics associated with bullies, research has shown them as angry, depressed, aggressive, hostile and domineering individuals showing high levels of externalizing and hyperactive behaviors with little fondness for school (Austin & Joseph, 1996; Bosworth, Espelage, & Simon, 1999; Byrne, 1994; Olweus, 1995; Rigby & Slee, 1991; Slee, 1994; Vaillancourt, Hymel, & McDougall, 2003) and high conflict within friendships. The typical bully has been found to be indiscriminately aggressive toward teachers, parents, siblings and peers. The question that stems from such researches is whether it holds true for both boy and girl bullies?

Gender and Bullying

Research on gender differences in bullying has examined the nature of inter-gender and in-gender bullying, and how boys and girls react to bullying and being bullied. Girls and boys display similar levels of bullying. However, boys report bullying more often and generally engage in overt physical forms of bullying, whereas girls tend to engage in covert psychological bullying (Crick, Grotpeter, & Bigbee, 2002; Hall, 1999; Pepler & Craig, 1997). Moretti (2002) reported that the gender gap in terms of seriousness and physical nature of bullying and violence is narrowing. A study by Peterson and Rigby (1999) found that although girls were involved as victims in less than half the amount of physical bullying than boys, boys were as involved as girls in the various forms of emotional/psychological bullying. Owens, Slee and Shute’s study (2000) found that girls were affected both by physical and indirect aggression. Of the studies done in India, Munni and Mahli (2006) reported that females were more likely to be victims of bullying. Kshirsagar, Agarwal, and Bavdekjar (2007), however, reported that the prevalence of bullying was the same among boys and girls in coeducation schools in India. A newspaper report of 2008, with reference to the BRITE (Bullying Research Initiative in Training and Education) study done at 12 English medium schools across North India, it was stated that 58.7 % boys in the age group of 14–18 felt that bullying was present on campus, whereas it was higher for girls, 65.09 %. While bullying among boys usually occurred through fights or use of abusive language, for girls, it took the form of teasing, name-calling or avoiding someone. Evidence also suggests that boys and girls bully equally and both can be targets. As a teen gets identified as a bully or victim, the peer contact contributes to perspective taking and understanding of self and others. With regard to this, the study also emphasized studying the peer relationships of those who bully.

Peer Relationships and Bullying

In middle childhood, the peers become an increasingly important context and component of development. Peer contact contributes to perspective taking and understanding of self and others. These changes enhance and impact peer interaction to make it more prosocial over the years. As a result, aggression declines, but the drop is greatest for physical attacks (Tremblay, 2000). However, this transition may not adopt the normal course for many. Children form peer groups, which are collectives that generate unique values and standards for behavior and a social structure of leaders and followers. According to Redl (1966), the beginning of peer group ties is also the time when some of the “nicest children begin to behave in the most awful way.” Studies have found that right from the third grade, while relational aggression is on a rise among girls, boys express their hostility in a more straightforward way in the form of verbal insults and pranks (Crick & Grotpeter, 1995). The concept of friendship too becomes a significant one during this stage. Yet the impact that friendships have on children depends on the nature of those friendships. The combination of being rejected by peers and being aggressive leads to various kinds of problems (Ladd, 2006; Rubin et al., 2006), one of them being peer victimization, in which certain children become frequent targets of verbal and physical attacks or other forms of abuse.

Aggression and Bullying

Researchers have identified many behavioral responses of high levels of anger in adolescence (Dodge & Coie, 1987). And most medical and social scientists agree that anger often serves as precursor to aggression and violence (Hinshaw et al., 1993; Walker et al., 1991). Aggression has been defined as a behavior directed toward the goal of harming another living being, who is motivated to avoid such treatment. Research has identified several etiological factors that play a role in the development of anger leading to aggressive antisocial behaviors. These are: (1) School performance and development of anger: Tremblay (2000) reported poor academic achievement as a significant variable in early disruptive behavior and subsequent personality disorders in adolescence and adulthood. (2) Deficient cognitive processing and peer rejection: Some adolescents due to their reactive nature have difficulty in assessing and utilizing social cues, they misattribute peer interactions resulting in negative and hostile behaviors (Crick & Dodge, 1994) showed that excessively angered and aggressive youth display deficits in their cognitive processing of social cues. (3) Dysfunctional home environment and anger development: According to Huston (1991), several types of difficult family environments may cause emotional instability and increased anger and aggression among children and adolescents. In support, the social learning theory perspective, given by Bandura (1977), says that humans acquire aggressive responses through direct experience or by observing other’s behaviors. (4) Genetics and anger development: Over time, various studies have posited the role of the evolution in anger. (e.g., Hilton, Harris and Rice, 2000). Neuroanatomical studies (e.g., Shapiro & Hynd, 1993) show a direct relationship between antisocial behaviors and dysfunctions in the workings of the frontal lobe area. Further, biochemical studies (e.g., Rogeness, Javors and Pliszka, 1992) have identified the importance of neurotransmitters, neuro-hormones (cortisol and testosterone) in the regulation of increased aggressive behavior patterns.

Besides these factors, a more recent approach, called the General Aggression Model (Anderson, 1997; Anderson & Bushman, 2002), posits that both the current situation and those relating to the person play a role in initiation of anger. In addition to the belief that bullies are more aggressive than others, it is also thought that the anger is a defense they use for the lack of self-esteem.

Self-Esteem and Bullying

In a general sense, self-esteem might mean anything from the good feeling about oneself, freedom to choose, to total acceptance of oneself and living by one’s own values and convictions. Branden (1992) has defined self-esteem as a confidence in our ability to think, to cope with the basic challenges of life and confidence in our right to be successful and happy. According to Baumeister, Bushman and Campbell (2000), not many studies have found any link between self-esteem and aggression. The literature on self-esteem in relation to children who bully others is controversial. In a study done by O’Moore (2000), it was found that the more frequently the children were victimized or bullied others, the lower was their self-esteem. The typology and frequency of bullying and the age of the children when they were involved in bullying too influenced the status of the specific domains of self-esteem. The contention that bullies too have low self-esteem has been a debatable one. While studies have shown that students who report high levels of being victimized are relatively low in self-esteem (Olweus, 1993; Rigby & Slee, 1993), a question that has arisen is whether low self-esteem is a cause rather than an effect of being victimized. Contrary to the belief that bullies may have low self-esteem, Olweus (1993) has asserted that they tend to be average or high as far as self-esteem is concerned. This has implications for bullying intervention programs. A number of school and treatment intervention programmes that focus upon enhancing the self-esteem of children who bully may in turn end up creating more confident bullies (Limber, Nation, Tracy, Melton, & Flerx 2004) and so need to be reviewed.

Measures

The four tools used to collect the data were:

  1. 1.

    Peer Relations Questionnaire (PRQ) (Rigby & Slee 1993 ): It is a short 20-item questionnaire, developed by Rigby and Slee (1993). Suitable for students aged 8–18 years, it assesses relationships of an adolescent in terms of his bullying, victim and prosocial behavior. Higher scores on each of the scales implied that the child was more inclined to be a bully, victim or helpful, respectively. The PRQ is an internationally and frequently used self-report measure of bullying (Griffin & Gross, 2004, cited by Hulsey, 2005). Hulsey (2005) in her study found out that PRQ had a test–retest reliability of 69 % correspondence and moderate reliability for middle school (69.4 % correspondence).

  2. 2.

    Adolescent Anger Rating Scale (De Anna McKinnie Burney & Wheeler, 2008: The Adolescent Anger Rating Scale (AARS): It gathers information from adolescent aged 11 through 19 years. The items identify an adolescent’s typical mode of anger expression and Anger Control. The scale yields the score on the following subscales: (1) Instrumental Anger (IA) expressed as delayed or covert anger; (2) Reactive Anger (RA) expressed as overt anger; and (3) Anger Control (AC) expressed as proactive behavior to resolve anger responses. The AARS uses the IA, RA and AC as measures of specific aspects of anger in adolescents. These three subscales are also used to assess an adolescents’ overall expression of anger. The AARS is a well-standardized instrument with an internal consistency of the entire sample ranging from .81 to .92; the test–retest reliability as measured within a 2-week interval found the correlations ranging from .71 to .79. High positive correlations for convergent validity were observed between the subscales of Conners—Wells Self-Report Scales—long (Conners, 1997) and the AARS subscales ranging from .35 to .61

  3. 3.

    Rosenberg Self-Esteem Scale (Rosenberg, 1965 ): A measure of self-esteem, RSE, is a 10-question scale which offers four response choices, ranging from Strongly Agree to Strongly Disagree. A higher score implies higher or normal self-esteem. Studies showing the scale to be a valid and reliable unidimensional measure of self-esteem have found the reproducibility to be .92 and scalability to be .72 (Rosenberg, 1965); a 2-week test–retest coefficient of .85 (Silbert & Tipett, 1965). The convergent validity with Coopersmith Self-Esteem Inventory (Coopersmith, 1967) was found out to be .60 (Crandal, 1973).

  4. 4.

    Interview (Structured and Semi-structured): Interview is almost an infinitely flexible tool (Breakwell, 2006).The methods employed in the present study ranged from structured to semi-structured interview. A structured interview is a quantitative research method commonly employed in survey research. The aim of this approach is to ensure that each interviewee is presented with exactly the same questions in the same order. This ensures that answers can be reliably aggregated and that comparisons can be made with confidence between sample subgroups or between different survey periods.

A teacher-report is typically used as a supplement to other measures as teachers are not always aware of all aggressive behavior that occur between students, but are able to provide valuable information on the climates of their schools and classrooms (Cornell & Brockenbrough, 2004). Keeping this in mind, the teacher of each class was interviewed. An interview schedule was formulated before hand to include all the questions which would be relevant to the issue and purpose of the study. A total of eight questions were outlined. The focus of the questions was on seeking the information from the teacher in terms of the most and least popular child in the class, those who were academically bright, the naughtiest, received punishment, resorted to violence and many other related factors along with the description of each child.

Unlike a structured interview, a semi-structured one allows a respondent the time and scope to express his/her opinions or share feelings in response to each question. The focus of the interview is decided based on the area the researcher is interested in exploring. The objective is to understand the respondent’s point of view rather than making generalizations about the behavior. Semi-structured interviews were conducted with a few students randomly selected to represent those identified as “highly inclined to be bully” after having considered their scores on PRQ, AARS, and RSE and the teacher reports. These interviews with the students were an attempt to further understand those identified as “highly inclined to be bullies” as also to supplement the quantitative/objective assessment done with self-report measures. The interview schedule was planned before hand to include the questions pertaining to the issues to be investigated. These were the child’s description of him/herself, his relationship with peers, teachers and family members and (s)he being hurt by someone or having hurt someone. Though the direction of the interview was somewhat pre-decided, it gave ample flexibility to follow-up each participant’s leads.

Methodology

The research was conducted in a central school, where English is the medium of instruction and which caters to children from diverse backgrounds. Middle school (class 6th to 8th) stage being the commencement of early adolescence was considered an appropriate sample group for the study, with participants aged 12–14 years. Following a pilot study, the random sampling approach determined the initial sample size which constituted one section each from classes 6th, 7th and 8th. The students were then administered PRQ, in English, to identify students “highly inclined to be bully.” The terms “potential bully” and “highly inclined to be bully” have been used throughout the research owing to the fact that the PRQ does not categorizes students as bullies but identifies those who have a potential to become a bully. Thus, the term was used, after verifying with Rigby (author of PRQ). All the instructions and items were read out aloud to ensure that the students understood them. Some of the difficult words were translated in Hindi with the help of an English–Hindi dictionary to maintain authenticity and uniformity (as/if and when a query was posed). In order to derive the target sample, from the initial sample of 137 students, the scores obtained by the participants on the Bully scale of PRQ were statistically evaluated. The measure of central tendency considered was the median. Since the median refers to the mid-point, half of the sample (those above the median) could be put in the bracket of “inclined to be bully.” However, to increase the objectivity of the data and reduce the sample size, an upper quartile score for the boys and for the girls were located. Students whose scores were above the upper quartile were then marked as “highly inclined to be a bully.” This method of identifying students “highly inclined to be bully” was agreed upon through a personal communication that the researcher had with Rigby (author of PRQ). The target sample was narrowed to 45 students “highly inclined to bully.” Structured interviews were conducted with the class teachers to seek an understanding of how they identified the potential bullies among their students. This was accompanied by interviewing six students who were representative of the target group, using a semi-structured interview format in order to understand the factors that led them to bully others.

The research determined the prevalence and kind of bullying behavior, the anger levels, the self-esteem and peer relations with respect to bullying behavior according to each gender. Correlations between the variables along with t test at 95 and 99 % confidence level were used to establish or negate a significant relationship between various factors.

Results

The initial sample consisted of 137 students on whom the PRQ was administered. The class-wise means of boys and girls on the three scales of PRQ (Bully scale, Victim scale and Prosocial scale) are depicted in Table 1.

Table 1 Class-wise mean of boys and girls on the three scales of PRQ
Full size table

The findings suggest that it is boys who more often engage in bullying behavior as compared to girls, and this pattern is seen irrespective of the classes they are in. On the other hand, they are likely to be victims as well. The statistical analysis for the combined data, boys and girls data on the three scales of PRQ is given in Tables 2, 3 and 4, respectively.

Table 2 Statistical analysis for scores on three scales of PRQ (boys + girls)
Full size table
Table 3 Statistical analysis for scores on three scales of PRQ (boys)
Full size table
Table 4 Statistical analysis for scores on three scales of PRQ (girls)
Full size table

From the above tables, it can be seen that boys have a greater likelihood than girls to engage in bullying behavior, as is clearly evident when the means of boys and girls on the Bully scale are compared. Consequently, the score demarcating upper quartile of boys on the Bully scale is higher than that of the girls (12 vs. 8)*. Hereafter, the analysis of the results was done only for the target sample (45 students).

The mean of the 45 students’ scores on the three subscales of the PRQ is graphically represented in Fig. 1. The graph clearly shows a significant difference between the means of boys and girls on the Bully Scale. Differences in the means on the Victim scale are also evident between the boys and girls. The scores on the Bully scales are higher than the scores on the Victim scale for both boys and girls (group-wise or combined). However, girls tend to be slightly higher than boys on the Prosocial scale.

Fig. 1
figure 1

Means of the scores on the three scales of PRQ for the target Sample {N = 45 (25 + 20)}

Full size image

The target sample was required to fill the Adolescent Anger Rating Scale (AARS) and the Rosenberg Self-Esteem Scale (RSE). The AARS provided the students’ scores on Instrumental Anger, Reactive Anger and Anger Control as well as a Total Anger score derived with the help of a formula given. The RSE yielded scores indicating the level of self-esteem of the students. Figure 2 depicts the means of the scores on AARS and RSE for boys, girls and combined data.

Fig. 2
figure 2

Means of the scores on AARS and RSE for the target sample {n = 45 (25 + 20)}

Full size image

The data presented in the graph suggest that the mean of boys (88.56) on Total Anger score is greater than the mean of girls (80.30). This indicates that out of the students who are “highly inclined to be bully,” boys experience more anger than girls. On the other hand, a look at the bar column for RSE shows that the mean RSE score for girls (20.10) is slightly higher than that of boys (18.28). This suggests that girls have favorable self-esteem relative to boys.

In order to understand the types of aggressive (anger) behaviors that boys and girls indulge in, the mean of each subscale of the Adolescent Anger Rating Scale i.e., Instrumental Anger (IA), Reactive Anger (RA) and Anger Control (AC) were obtained (Fig. 3).

Fig. 3
figure 3

Means of the scores on the three subscales and the Total Anger of AARS for the target sample {n = 45 (25 + 20)}

Full size image

The data shown in Fig. 3 suggest that boys have the higher means in all the three subscales i.e., IA, RA and AC as compared to those of girls. This suggests that boys perhaps engage in both the forms of anger/aggression (instrumental and reactive) as also attempt in Anger Control more than the girls. However, when compared across the scales, both boys and girls engage more in Instrumental Anger than in Reactive Anger.

Further analyses of the data from the target sample concerned the relationship between all the variables. Table 5 shows the correlation between the three scales of PRQ.

Table 5 Correlation between the scores on the three scales of PRQ for boys, girls and combined data
Full size table

It is evident from Table 5 that there exists a significant correlation between the Bully score and the Victim score for the combined data, which means that in the sample regardless of gender, a student who is ‘highly inclined to be a bully’ is also quite likely to be a victim as well. A negative correlation between the Victim and the Prosocial score for girls and boys suggests that those who are victimized are less likely to engage in prosocial behavior. However, this relationship is not a statistically significant one.

The correlation was also found between the self-esteem score (as measured by RSE inventory) the Total Anger score (as measured by AARS) and the Bully score (as measured by PRQ) for boys, girls and the combined data (boys + girls) (Table 6).

Table 6 Correlations between the scores on the Bully scale, RSE and the Total Anger AARS for boys, girls and combined data
Full size table

Significant positive correlation can be seen between the Bully score and the Total Anger score both for boys as well as the combined data (of boys and girls). The positive correlation suggests that if a child is a bully, the more likely it is that he will be high on anger. For girls, the relationship is not significant.

The t test yielding differences between the means of boys and girls for the Bully score, Victim score and the Prosocial score was employed (Table 7).

Table 7 t test values indicating the difference between means for boys and girls on the three scales of PRQ
Full size table

As shown in Table 7, the t test values suggest that there is a significant difference between the means of boys and girls for both the Bully scale score and the Victim scale score, respectively. The t test values on the Prosocial score, however, suggest that there is no significant difference between the means of boys and girls. Thus, the significance of the t test values supports the differences in the means as depicted in Fig. 1.

Along with the t test values for the three scales of PRQ, the t test was done for Self-Esteem scores and the Total Anger scores across the two groups of boys and girls (Table 8).

Table 8 t test values indicating the difference between means for boys and girls on Self-esteem score and Total Anger score
Full size table

The t test values on RSE score and Total Anger score (Table 8) indicate significant difference between the means of boys and girls on Total Anger score (p < .05). This difference validates the results depicted in Fig. 2.

Results from the Structured Interviews with Teachers

The teachers indicated the names of the “potential bullies.” They identified only boys who were: least popular, outspoken and loud, often punished, least punctual, average or below average in academics, used verbal and physical violence, respected teachers but did not obey them and those who were friendly with classmates. The names given by the teachers of students inclined to be bullies matched the results found using PRQ, but not in the case of girls.

Results from the Semi-structured Interviews with Students

The themes reflected in the interviews with students indicated that boys were more likely to aggress (verbally/physically) when provoked or abused. Girls, on the other hand, aggressed (covertly) when their reputation or image was attacked or when they were emotionally hurt in a relationship. However, the instances of being punished at home physically were similar across the two genders.

Discussion

The aim of the present study was to come closer to an understanding of the potential bully. In this regard, an attempt has been made to do so by discussing the results obtained from the quantitative and the qualitative data.

The findings (as depicted in Tables 1, 2, 3, 4) for the initial data as well as for the target sample (Fig. 1) suggest that boys seem to be engaging in bullying behavior more than the girls. Research shows that physical abuse tends to occur more often among boys than girls at all educational levels (e.g., elementary, high school, college) (Smith et al., 2005; Chapell et al., 2006; Espelage & Swearer, 2003). In the responses given by the teachers, it was observed that they recalled the names of boys more readily than the names of girls as bullies. In fact, it seems that even boys who are bullies, too, may tend to overlook girls as involved in bullying. One of the boys interviewed in the present study said “Only 2–3 boys in the class are good. All girls are good in class except boys.” These findings validate the significant difference (p < 0.01) between the means of the scores of boys and girls on the Bully scale as seen from the t value (Table 7). This may be because of the manner in which boys bully is more direct and explicit, thereby making it obvious that physical harm or injury has been caused. In contrast, the way in which girls bully is more subtle and covert and so the psychological bullying they indulge in is not recognized as such.

Many reasons may be put forth to explain boys being overtly expressive in their bullying behavior. One might be parents, who serve as powerful role models. Aggressive and violent behavior of parents may lead children to believe that violence is acceptable and validated and so can be used against peers when angry. Wilson in her article (2006) reported that children who are bullies at school are more likely to have witnessed violence at home compared to children who did not see this aggressive behavior. In the present study, all the students interviewed (regardless of gender) when asked about their parent’s reaction to something not approved by them, responded by saying that they were either scolded or beaten up by their parents. Like with boys, girls might emulate their mothers when it comes to expressing their anger covertly. According to Crick and Grotpeter (1996), girls are subtle in their expression as they engage in relational bullying. The relational strategies are utilized in female bullying because as girls enter adolescence, they substantially invest in social comparisons and peer acceptance for self-worth, making them particularly susceptible to, and highly aware of, the impressions of others (Casey-Cannon, Hayward, & Gowen, 2001). Therefore, the most effective way to harm a girl is to manipulate her relationships within her peer group, which is precisely the aim of relational aggression (Goodwin, 2002). A girl interviewed in this study, when asked if she had ever hurt anyone, responded by saying that “At my old home I had a friend. She was younger to us. She would say anything unnecessarily. And we had a rule that elder friends are to be addressed as “didi”, which she did not follow. So we stopped talking to her”.

In contrast to the studies which reveal that boys engage in physical bullying and girls in relational bullying, Peterson and Rigby (1999) found that boys were as involved as girls in the various forms of emotional/psychological bullying. Researchers (Moretti, 2002; Petersen & Rigby, 1999) have found that the difference between boys and girls in terms of intensity and physical nature of bullying and violence was narrowing. Glimpses of this finding could be seen in our study as well. For example, one of the girls interviewed said “My brother was troubling me when I was watching TV by coming again and again in front of the television, I tried to slap him, but by mistake it hit badly in his eye, and he started crying. I felt bad.” On the other hand, a boy who was interviewed said “They wrote something very disgusting, insulting and cheap about my mother on the paper, and kept abusing me.” This suggests that both boys and girls have a potential to engage in physical as well as relational bullying. Hence, the results reveal that gender differences exist with regard to bullying behavior with boys engaging significantly more than girls.

With respect to the anger levels, the findings of the current study revealed a clear difference between the means of Total Anger score between boys (88.56) and girls (80.30) (See Fig. 2). These data were further validated by the t test value and significant difference (p < .05) between boys and girls on the Total Anger score (Table 8). It suggested that boys have higher anger levels as compared to girls. On studying the types of aggression, it was found that though boys scored higher than girls on both Instrumental and Reactive Anger, but they indulged more in Instrumental than Reactive Anger (Fig. 3). Thus, the high score of boys and girls on Anstrumental Anger probably suggests simply the type of anger they resort to. Manifestation of anger is a different aspect, as suggested by the kind of bullying behavior they indulge in. In the present study, when the students found “highly inclined to bully” were asked if they got angry, some of the responses gathered were:

  • “Sometimes yes, when others abuse me; when they hit me with ball in the recess. I sometimes start fighting or tell ma’am”.

  • “I get angry when someone says that I’m bad, I don’t know anything or I don’t respect others”. “When my friend instigates others not to talk to me that too makes me angry”.

Thus, potential bullies may get angry when they are provoked or when their reputation is being damaged. Besides these, there might be many factors that could contribute to the aggression levels of a child who bullies others. One of the prime factors might be the child’s family environment. Some of the bullying kids may come from the families where they do not get adequate emotional support. Studies have found that childhood experiences with aggression, such as physical discipline (e.g., spanking, inconsistent punishment, family violence, victimization by siblings, paternal bullying), are positively associated with bullying (Espalage, Bosworth, & Simon, 2000; Farrington, 1993; Strassberg, Dodge, Pettit, & Bates, 1994). Some of the responses which were seen, in the interview with children, in the present study were-

  • “Mother is stricter. She slaps me when my work is not complete and when someone fights with me”.

  • “My parents scold me and beat me also. If I have done something wrong, it is expected I will get a scolding”.

Inclusive of the above examples, it was observed that all of the students interviewed by the researcher reported being scolded or beaten up by their parents or other significant persons whenever they did something that was unacceptable or considered inappropriate. Thus, families, where there is a lack of warmth and involvement on the part of parents, where the parents are overly permissive, where there is a lack of supervision by parents, or where parents “rule their children by stick” using harsh, physical discipline, may contribute in the making of bullies.

As adolescence is a time when peers become an integral part, peer interaction has a significant role to play in a child’s bullying behavior. Cross-sectional research on the adjustment of bullies revealed that bullies are more rejected and less popular (Boulton & Smith, 1994) and display more antisocial, aggressive and disruptive behavior than non-involved children (Pellegrini, Bartini & Brooks, 1999; Rigby & Cox, 1996). One of the boys interviewed in the present research in his experience of being hurt said “He (other boy) beats everyone. He makes up stories and complains to ma’am about me. No one in the class talks to that child.” Consequently, this indicates that the other boy who was a bully seemed to be less popular and was rejected by his classmates, as gathered by the response of the boy interviewed. Hence, the need to dominate one’s peers and gain social approval may be one of the other factors leading a child to engage in bullying behavior.

Another factor contributing to the child’s aggression perhaps could be his/her own experience of being a victim. According to Schwartz (2000), some children have a tendency to be involved in bullying, as a bully, a victim, or a bully–victim. A bully–victim is a child who has been victimized by someone and goes on to bully someone else, so as to vent out the suppressed anger she/he carries against their assailant. The desire to overcome the sense of having been hurt and meek finds some compensation by bullying others. As shown in Fig. 1, though the means of boys on Bully scale is higher in comparison with the girls, this is also accompanied by a high Victim score of boys than that of girls.

Therefore, with reference to the studies, the quantitative (Tables 3, 4, 5, 6, 7, 8; Figs. 1, 2, 3) and the qualitative results, it can be said that there are various factors which lead a potential bully to engage in aggressive acts, which might be either physical or verbal. And there is no one single factor which is solely responsible for the aggression level of a bully. A bully’s family environment, peer relations and past experience as a victim all have a role to play in the anger expressed by him/her.

Another important aspect related to bullying which has been debated is the level of self-esteem of bullies. The current study used Rosenberg Self-Esteem Inventory (RSE) to assess the same. As per the RSE, the range of the raw scores is from 10 to 40, with a score of 14 and below indicating low self-esteem. In the present study, the raw scores ranged from 13 to 26, with only 3 students showing low self-esteem and the rest of the sample having normal self-esteem. Figure 2 shows that the girl’s mean score (20.10) on RSE is slightly higher than that of the boys (18.28), while the self-esteem of the combined data is 19.09. Further, Table 6 suggests no significant correlation between the self-esteem, Bully score and the Total Anger score for both boys and girls. Additionally, Table 8 indicates no significant difference between the means of boys and girls on RSE score. This suggests that there were no significant differences among those “highly inclined to be bully” with regard to self-esteem, thus implying that potential bullies do not have low self-esteem.

There might be many reasons for there being no gender differences, as far as self-esteem is concerned. Studies have found that although individual differences exist, during childhood and adolescence, perceived physical appearance correlates more strongly with overall self-worth than any other self-esteem factor (Hymel, Tarulli, Hayden Thomson, & Terrell-Deutsch, 1999). According to Twenge and Campbell (2001), self-esteem is on the rise and remains high for majority of young people. The self-esteem of the students in the target sample may have been influenced by other factors such as academic performance, praise and appreciation by teachers, excellence in extra-curricular activities or popularity in class among others, those not emphasized in the present study and need further understanding in future research projects.

With regard to the relationship between the different variables, referring to Fig. 1, when the means of boys and girls on the Bully scale and Victim scale are compared, it can be seen that the difference between the girls being a bully and a victim is far less than that seen for boys. Research findings suggest that boys are more likely to both bully and be bullied than girls (Baldry & Farrington, 2000; Garandeau & Cillessen, 2006; Liang, Flisher & Lombard, 2007). This perhaps indicates that some boys might get an opportunity to externalize their anger thereby taking the role of a bully, while there are those who continue to be victimized by their peers. Hence, bullying can also have a “domino effect.” A child who is continuously being bullied may be unable to get back directly but may be motivated to bully someone else. Unlike boys, girls may find refuge in covert or relational bullying to victimize someone else in lieu of their perpetrator. Additionally, girls being placed in roles requiring conciliatory and accepting behavior may withdraw, thus taking on role of victims. This may be one plausible explanation for a lesser difference between the mean of girls on the Bully and the Victim scale in our study. However, on combining the data of girls and boys, there appears significant correlation between the Bully and the Victim scale (Table 5). This supports Evelyn Field’s (2007) work where she mentions that boys and girls can both equally be bully and targets as well.

The study also found a significant relationship to exist between the Bully score and the Total Anger score. Significant positive correlation (p < 0.01) was found between the Bully score and the Total Anger score for both boys and the combined data (of boys and girls) (Table 5). This is to say that the more a child is predisposed to bullying behavior, the more likely it is that he/she will show high levels of anger as well. With reference to Fig. 2, there exists a clear difference between the means of Total Anger score between boys (88.56) and girls (80.30). This is further validated by the t test indicating significant difference (p < .05) between boys and girls on the Total Anger score (Table 8). According to Pearce and Thompson (1998), the typical bully is indiscriminately aggressive toward teachers, parents, siblings and peers. She/he usually dislikes and has not adjusted to school, has poor impulse control, wishes to dominate, is physically and emotionally strong, craves social prestige and is insensitive to the feelings of others. All of these behavior characteristics of a bully were somewhere reflected in the quantitative as well as the qualitative responses gathered from the teachers as well as the students.

The current study also employed structured interviews with teachers to ascertain the level of congruence between the assessment of the children “inclined to be bullies” and the students’ self-reports. Results revealed teachers to be sensitive to direct bullying behavior, identifying boys as bullies and overlooking girls. This suggests that the teacher’s biases with respect to girls, to the disadvantage of boys, impacts their identification of the potential bullies. Further, relational bullying being covert and hard to detect made it difficult for untrained teachers to recognize it. Therefore, bullying can go unnoticed until long after an incident has occurred, even in cases of frequent victimization (Casey-Cannon et al., 2001). Hence, our study seems to support the observations made by teachers interviewed.

Of the responses gathered from interviews with students, it could be seen that boys were more likely to be victims turned into bullies, owing to their victimization caused either physically or emotionally; and the bullying behavior they engaged in was often physical and verbal. In contrast, girls seemed to engage in relational bullying (by avoiding a particular classmate or by not talking to another girl). However, examining responses of boys and girls, both were equally likely to engage in physical as well as covert bullying. As far as peer relations were concerned, potential bully (girls) seemed to have cordial relations with their classmates in contrast to potential bully (boys) who were mostly engaged in interpersonal conflict. Thus, the responses of students suggested that girls’ engagement primarily in covert bullying would be subtle and almost invisible, though in no way less injurious. However, boys would engage in direct forms of bullying, adversely affecting their evaluations made by the teachers, classmates and peers.

Conclusion

In summary, it was found that there existed gender differences with regard to bullying behavior and anger levels. With respect to self-esteem, no significant differences between boys and girls were found. Qualitative data from the teachers reasonably matched with that of the students’ identified as “highly inclined to be bully” via self-reports. This research identified students as “inclined to be bully” not to label them, but with a view that their recognition would lead to timely intervention and support measures by the school community. On having come closer to the potential bully, the researcher observed that more often than not he is a boy, though girls also show the potential to do so, albeit, in a different manner.

The research, through its findings, sensitizes the entire school and parent community along with the bully, victim and the bystanders to the misconceptions of a “potential bully” and their behavior. It brings to light the fact that bullying behavior can be exhibited regardless of one’s gender. Further, terms such as “bullying” and “aggression” are often considered as synonyms. The study provides clarity to the student, teacher, parent community and mental health professional regarding the same.

It emphasizes the need for comprehensive school-based interventions wherein the role of teachers, school management and mental health professionals is integrated with the engagement of family to address the “bullying behavior” at a larger level. At the management level, it is necessary for interventions to focus on the policies of the school and modifying them so as to create a “zero tolerance for bullying” environment. This requires involvement of student representatives, administrators, parents, teachers and community members. Once the policy is in place, further steps include creating awareness about bullying behavior through the use of role plays, dramas, workshops and sensitizing the children about the consequences of bullying and victimization.

Of those identified as “potential bullies,” the school authorities need to hone their talents and engage them in psychotherapy with school counselors, psychologists. These interventions must focus on enhancing peer relationships, developing life skills and managing anger via workshops, group work and enactments.

As reflected in the findings, the inability of teachers to identify the potential bullies calls for the need for specialized training programs. These involve building and enhancing teacher sensitivity and competence to recognize early signs of bullying and take required steps at the time of crisis.

Since the present study was limited to one school, the findings, even though new, restrict the generalization to the other public or government schools. Future research could compare between a public versus government or a coed versus a single boys’ or girls’ school in terms of the bullying behavior prevalent. A larger sample could be taken so as to provide additional insights. In addition, perhaps a control group could be used to understand comparison and bullying behavior patterns. While a few controls and variables were employed in the current study, prospective researches could study variables such as the personality, coping mechanisms, well-being and general health with controls such as SES and family background for instance. Lastly, bullying as an issue also involves the victim, the bully–victims and the bystanders. Research could also be attempted to understand the victim’s perspective along with the interface or transformation of victim–bully or bully–victim.