A cross-sectional survey was conducted in October and November of 2014 in five conveniently selected English medium schools of Vadodara (2) and Anand (3) after approvals from the Institutional Ethics Committee of our institute, and school principals. Survey was administered to participants after obtaining a written informed assent.

Survey included demographic variables like age, sex, height and weight, scholastic performance and number of friends, Peer Interaction in Primary School Questionnaire (PIPS) [6] and Strengths and Difficulties Questionnaire (SDQ) [7].

English self-report version (11-17y) of SDQ was used to assess psychosocial consequences in terms of emotional, behavioral and inter-personal difficulties. SDQ has 25 items, both positively and negatively worded with 5 subscales. Singh, et al. [8] have reported acceptable confirmatory factor analysis properties of English self-report SDQ in Indian adolescents. Segregation into victim, bully and bully-victim was done based on two SDQ questions viz. "Other people or young children pick on me or bully me" and "I fight a lot. I can make other people do what I want."

PIPS is a self-reporting bullying and victimization measure consisting of 22 items scored on a three-point scale: "never", "sometimes" and "a lot". It has two subscales viz. Victim and Bully. PIPS has been designed for use in primary schools; however, the behaviors listed were equally relevant in current context [6].

PIPS subscale scores were used to assess variance by demographic categories. Means on individual questions of the PIPS were used to identify most frequent bully behaviors and victim experiences. SDQ-based categorization was used for calculation of prevalence and evaluation of differences in psychosocial consequences.

PIPS subscale scores were significantly higher in victims (n=328) compared to non-victims (n=778) [5.64 vs. 3.34, P <0.001] and bullies (n=331) compared to non-bullies (n=775) [4.10 vs. 2.73, P<0.001]. Overall prevalence of any form of bullying was 49%, with bullies being 29.9% and victims being 29.7%. Being teased (0.68) and made fun of (0.63) were the most frequently reported experiences by victims. Bullies reported Teasing (0.75) and making fun of others (0.45) as the most frequently used behaviors (Web Table I). PIPS victim scale mean score was higher in the victim and bully-victim groups but, equivocal in the bully group. Mean PIPS bully scale score was higher in the bully-victim group but was almost equal in the bully and victim group.

Males reported significantly higher bullying experiences (both bullying others and victim experiences). Age had no association with either being a bully (P=0.07) or being a victim (P=0.37). Overweight/obese (P=0.02) and students having less friends (P=0.001) were more likely to be victims (Table I).

TABLE I Differences in Bullies and Victims by Demographic Characteristics

Victim and bully-victim groups had higher total difficulties score compared to the non-affected group on SDQ (P<0.001). Victim group had higher scores on emotional problems, hyperactivity and peer problems compared to Bully group which had higher scores on conduct problems, hyperactivity and lower scores on prosocial behavior. Bully-victim group was most severely affected with higher scores on all subscales of SDQ as compared to the non-affected group.

We found 29.7% students being victimized by bullying. Girls who are victims reported higher experience of emotional and sensitive forms of bullying whereas boys who are victims reported higher experience of physical and verbal means of bullying. Psycho-social consequences assessment using SDQ showed that victims had higher emotional problems, hyperactivity and peer problems.

Earlier researchers [1,9,10] have identified similar patterns and have speculated that some victims of bullying can be timid and may find it difficult to take part in relationships with peers. Whereas, those who are hyperactive may exhibit a combination of anxiety and aggressive traits (more likely to get categorized as bully-victim) that may provoke their peers with hyperactive and irritating behavior [1,9].

The strengths of the current study is its large sample size across two centers. The limitations of the study were lack of prior validity of PIPS in the study population, a sampling strategy of convenience, inclusion of only urban English medium schools. Further, link between computer gaming involving virtual violence and bullying behavior was not assessed.

Bullying-related involvement in our study is comparable to recent Indian studies [3,5] but quite high compared to those reported in Western literature. Health Behavior in School-Aged Children survey conducted in 40 countries documented the prevalence of bullies, victims and bully-victims to be 10.7%, 12.6% and 3.6%, respectively [11]. Like earlier studies, verbal bullying (teasing and making fun) is the most frequent with physical forms not being infrequent [2,3,5]. Previous research showed that boys are more likely to be bullies and girls are more likely to be victims [3,12]. In our study, boys are predominant in both bullies and victim sub-groups.

In summary, the fact that prevalence of bullying is high combined with previous knowledge that bullying peaks in middle school years [12], has short-term [2,3], long-term consequences [4], and that parents and teachers may be many a times unaware [2] of the prevalence, raise serious concern to address this problem. Role of schools in bullying prevention in Indian context needs to be explored further and appropriate guidelines needs to be developed.

Contributors: HP and JV: Study-design, data acquisition, data analysis, writing the manuscript. SN: Design and planning of the study, data acquisition, data analysis, revision of the manuscript for important intellectual content. SS: Data acquisition, data analysis, and revision of the manuscript for important intellectual content. AP: Study design, data analysis, and revision of the manuscript for important intellectual content. All authors approved the final manuscript.

Funding: None; Competing Interest: None stated.