The causes of bullying: results from the National Survey of School Health (PeNSE)

Objective: to identify the characteristics and reasons reported by Brazilian students for school bullying. Method: this cross-sectional study uses data from an epidemiological survey (National Survey of School Health) conducted in 2012. A total of 109,104 9th grade students from private and public schools participated. Data were collected through a self-applied questionnaire and the analysis was performed using SPSS, version 20, Complex Samples Module. Results: the prevalence of bullying was 7.2%, most frequently affecting Afro-descendant or indigenous younger boys, whose mothers were characterized by low levels of education. In regard to the reasons/causes of bullying, 51.2% did not specify; the second highest frequency of victimization was related to body appearance (18.6%); followed by facial appearance (16.2%); race/color (6.8%); sexual orientation 2.9%; religion 2.5%; and region of origin 1.7%. The results are similar to those found in other sociocultural contexts. Conclusion: the problem belongs to the health field because it gathers aspects that determine the students' health-disease-care continuum.


Introduction
The term bullying refers to a specific form of aggressive and violent behavior among peers in the school context. It is characterized by three criteria: intentionality, repeatability and imbalance of power (1) .
Given the emphasis of this definition, school bullying are acts that repeat over time and involve a desire to harm colleagues or expose them to negative situations, while those exposed to negative situations have difficulty defending themselves (1)(2) . This phenomenon may manifest directly and physically (e.g., hitting, spitting), verbally (derogatory nicknames, threats, insults, gossip), or through cyber-bullying (using social, electronic or communication media -internet, phone) or indirectly in situations where there is no direct confrontation among those involved (social exclusion, gossip) (3)(4) .
Bullying is acknowledged as a relationship problem in which power is claimed through the use of violence and is a reality among school-aged children and adolescents in different cultural contexts (4) and a severe problem in many countries (3)(4)(5) . This phenomenon may lead students to experience psychological distress, compromise the teaching-learning process and influence how individuals respond to social demands over the course of their lives. These negative consequences (4,6) , entailed for all those involved and associated with increased prevalence and frequency with which bullying occurs (7)(8) , transformed bullying into a severe public health problem worldwide (9)(10) .
Studies show that both boys and girls become involved in situations of violence at school, though the actions in which they engage are different. Boys are more likely to experience physical bullying, while girls engage in indirect or verbal exchanges (1,5,10) . Even though there are an increased number of studies addressing school bullying, few of them address causal factors or the reasons determining the phenomenon. In general, the focus of investigations is on the characteristics of the students involved, the phenomenon's variables and the nuances it assumes in the school context without, however, establishing the reasons that explain this phenomenon.
In this sense, evidence from the scientific literature addressing this subject suggests that the dynamics of bullying is a result of the students' characteristics, the vulnerability or social status of one student in relation to another, that differentiate and segregate peers (3) . A study conducted in Netherlands with 80,770 students reports that the reasons students presented for the practice of bullying were physical appearance, individual behavior, level of school performance, physical or mental disabilities, religious aspects, gender issues, sexual orientation, and the inappropriate manner some students dealt with punishment (11) . The average prevalence of students identified as involved in bullying was 32.5% (11) .
A longitudinal studied conducted in the United States reports empirical evidence of increased school bullying beginning in the second half of the 2000s, with a prevalence of 25.8% in 2009 (12) . The study reports bullying was more common and more intense among boys, Afro-descendants, from rural areas, living with single parents, with low school performance and a low level of religious identification (12) . A Swedish study, reporting a prevalence of 44% of victims and aggressors, reports that adolescents tend to explain the phenomenon in terms of individual reasons instead of offering other dimensions like peer groups, school context, or social issues (5) . The study also reveals that aggressors were more likely to blame the victim (5) .
In Brazil, the complexity of concrete problems such as bullying and a concern with school health culminated  (13) . Therefore, identifying the causes and reasons students become involved with bullying is essential to implementing coping strategies focused on human development and health promotion in the school context.
From this perspective, this study's aim was to identify the reasons associated with school bullying reported by adolescents in public and private schools in Brazil. and health protection in a sample of 8 th grade students attending daytime programs of public and private schools located in urban or rural areas from the entire Brazilian territory. The 9 th grade was chosen because it is the minimum level of education required to complete the self-administered questionnaire during data collection.

Study setting and sampling
The 2010 School Census was used to select the sample and those schools reporting 9 th grade classes administered during daytime hours were included in the list; nighttime programs were excluded. The The final sample included 109,104 students, i.e., 83% of those considered eligible for the study (14) .
A total of 86% of students in the sample surveyed in 2012 were between 13 and 15 years old; 47.8% were male and 52.2% were female; and 17.2% students were from private schools and 82.8% were from public schools (14) .

Procedures
Data were collected using smartphones, which were included in the structured, self-applied questionnaires with thematic modules that varied in the number of questions contained. Bullying was one of the dimensions addressed. Data collection was implemented by previously trained agents from the Brazilian Institute of Geography and Statistics (IBGE), in schools during classes from April to September 2012. Further details concerning the methodology can be obtained in specific publications (14) .

Studied variables
The variable bullying was obtained through the question: "How often did some of your friends belittle,

Statistical Analysis
The analysis was performed through the computation of the prevalence of the variables experiencing bullying and their respective confidence intervals of 95%, according to the sociodemographic aspects of experiencing bullying (sex, age, race/ color, religion, public or private school, mother's education). The reasons/causes of experiencing bullying reported by the students were analyzed according to sociodemographic aspects stratified by race or color, religion, facial appearance, body appearance, sexual orientation, region of origin, others.
When the reason one experienced bullying was reported to be the appearance of body, it was crosstabulated with the variable Body Image, which was verified by the question: In regard to your body, do you consider yourself: Too thin, Thin, Normal, Fat, Too fat?
These analyses were performed using SPSS, version 20, with the Complex Samples Module, appropriate for data analyses obtained by a complex sampling plan (15) .   Table 1 shows that 7.2% (CI95% 6.6-7.8) of the students reported having experienced bullying, always or almost always felt humiliated, by schoolmates. The percentages were higher among male students, 7.9% (CI95% 7.0-9.1), in comparison to female students, 6.5% (CI95% 6.2-6.7); among students whose mothers were characterized by low levels of education, 8 religion, 2.5% (CI95% 1.9-3.2); and region of origin, 1.7 (CI95%1.5-2). The frequencies of those reporting having experienced bullying and those reporting always or almost always experienced bullying in the last 30 days were similar, except for those reporting the reason was their race/color, among whom frequency increased to always, as shown in Table 2.

Results
Body appearance was cross-tabulated with the variable body image for those reporting that the appearance of their bodies was the reason they suffered bullying, which showed bullying was more frequent among those reporting they were either too fat or too thin, 19.2% (CI95% 15.1-24) and 12.1% (CI95% 10.4-14.0), respectively (Table 3).    Table 4.

Discussion
This study's findings show that 7.2% of the students experienced bullying, which was more frequently reported by younger boys, whose mothers present lower levels of education, and are of Afro-descent or indigenous. Most did not report the reason or cause that triggers bullying. In regard to differences between sexes, the causes reported by boys and girls were similar, mostly appearance of the face and body, however, boys This study highlights that "other reasons/causes" is the most frequent option chosen to explain bullying.
The frequency with which this option was chosen may be due to the poor understanding of students concerning the process of victimization or how they qualify jokes or the experience of being bullied. The process of victimization is characterized by receiving negative attention or aggressive behavior from peers over time and what determines its occurrence is being different or behaving differently others (2) . Investigating what causes the phenomenon based on self-reports addresses these dimensions and the sensitive nature of the issues implicated in the issue.
Almost a fifth of the students reported body appearance, followed by facial appearance, as being causes of bullying. Similar results were found in other contexts that indicate that physical appearance is one of the main reasons a student becomes a victim of bullying (16) . A potential interpretation for this information involves culturally valued social standards in which diversity and differences are not tolerated. One epidemiological study conducted with 1,230 students from a city in Rio Grande do Sul, Brazil, identified that 30.1% were overweight or obese, showing that students dissatisfied with their body image were three times more likely to be victims of bullying. Statistically, however, excess weight was not significantly associated with the phenomenon (16) . In turn one study, similar to this study, that was developed in Ireland reports that body image, such as considering oneself to be thin or too thin, was significantly associated with being a victim (17) .
Classical studies addressing this phenomenon do not report evidence that body image is a determinant factor in the process of victimization (1) . Other studies however, verify that victims often present characteristics that distinguish them from most of their peers, such as obesity, thinness, or the use of prosthetics or orthotics (18)(19) .
A student's skin color or race was also reported as being significantly associated with victimization. Afrodescendant students were four times more likely to experience bullying, while indigenous people were two times more likely to experience bullying. This dimension is also linked to social and cultural issues, to racism and prejudice, since there is a hegemonic pattern of valuing white skin (20)(21)(22)(23) . One study in the United States correlated race with gender and identified that these variables were significant predictors of bullying. The study shows that boys were 25.5% more likely to become victims than girls, while Afro-descendant students were 46.3% more likely to become victims at school than Caucasian students (23) . Afro-descendant and indigenous students addressed in this study were also more likely to become victims due to their race/color. It is worth noting that individuals of mixed race did not present the same rates of being bullied, an aspect that shows the importance of verifying whether students from different races have different criteria to identify and assess violent practices.
We cannot ignore the factors and individual variables that explain the phenomenon, as we cannot ignore contextual factors, such as mother's education, in the determination of bullying. As observed, the indication of no maternal education was the most prevalent for victimization and the scientific literature considers this variable to be a demographic predictor of students' success or failure at school. One study recently conducted in the United States reports that students whose mothers presented low levels of education were more likely to become victims (12) . In general, results concerning association between mother's education and involvement with bullying are explored because the mother's education is considered to be relevant within the families' set of social and cultural characteristics.
Other issues, such as the students' sexual orientation, religion and region of origin, are not shown to be expressive causes for victimization. In fact, these individual characteristics of students are less frequently observed than other characteristics. Nonetheless, they are manifested differently between sexes; for instance, boys more frequently report victimization associated with sexual orientation than do girls. Additionally, the literature shows that sexual orientation is one of the reasons related to bullying (11) . Therefore, these are important variables through which the phenomenon may be approached and related to proposing interventions intended to understand diversity, especially considering the diffuse nature of these in modern times and the emergence of other expressions of sexuality, religiosity, and migratory movements that require understanding and tolerance of diversity (4,(23)(24) . Overall

Conclusions
This study's results concerning the identification