SciELO - Scientific Electronic Library Online

 
vol.46 número161Repetência e indisciplina: evidências de Brasil e Portugal no Pisa 2012Indisciplina escolar: um itinerário de um tema/problema de pesquisa índice de autoresíndice de assuntospesquisa de artigos
Home Pagelista alfabética de periódicos  

Serviços Personalizados

Journal

Artigo

Compartilhar


Cadernos de Pesquisa

versão impressa ISSN 0100-1574versão On-line ISSN 1980-5314

Cad. Pesqui. vol.46 no.161 São Paulo jul./set. 2016  Epub 01-Jul-2016

https://doi.org/10.1590/198053143888 

ISSUE IN FOCUS

Oppression in schools: bullying among students in basic education

Bruno Lazzarotti Diniz CostaI 

Cíntia Santana e Silva

IEscola de Governo, Fundação João Pinheiro, Belo Horizonte, Minas Gerais, Brasil, bruno.diniz@fjp.mg.gov.br


Abstract

The article examines the phenomenon of bullying among students in basic education (elementary-middle school and high school) in Minas Gerais' public schools. The empirical analysis relies on data obtained from 5,300 students and 243 teachers during 2012 and 2013. The data has been analyzed based on general linear hierarchical models. The probability of incidence of bullying was estimated based on a two-level structure, one looking at individuals and one at schools. The findings that show the distribution of bullying are explained, on an individual level, by the quality of the bond between student and school and by the practice of deviant behaviors; and on the school level, by contextual characteristics such as the general sense of satisfaction, quality of teacher-student relationships and methods of conflict restraint.

Keywords: Bullying; Violence in Schools; Elementary Schools; Middle Schools; High Schools

Resumo

O artigo examina o fenômeno bullying entre os alunos do ensino básico das escolas estaduais de Minas Gerais. A análise empírica apoiou-se em dados obtidos junto a 5.300 alunos e 243 professores nos anos de 2012 e 2013. Os dados foram analisados com base em modelos lineares generalizados hierárquicos. Foi estimada a chance de ocorrência da prática de bullying a partir de uma estrutura de dois níveis, o dos alunos e o das escolas. Verificou-se que a distribuição do bullying é explicada, no nível individual, pela qualidade do vínculo do aluno com a escola e pela aderência a comportamentos desviantes; no nível escolar, por características contextuais, como clima geral de satisfação, qualidade da relação professor-aluno e método de contenção de conflitos.

Palavras-chve: Bullying; Escolas; Ensino Básico

Résumé

L'article examine le phénomène du harcèlement entre les élèves de l'enseignement primaire et secondaire des écoles de l'état de Minas Gerais. L'analyse empirique s'est reposée sur les données obtenues auprès de 5.300 élèves et 243 enseignants dans les années de 2012 et 2013. les données ont été analysées ayant pour base des modèles linéaires généralisés hiérarchiques. La probabilité d'occurrence de harcèlement a été a été estimée à partir d'une structure de deux niveaux, celui des élèves et celui de écoles. On a vérifié que la distribution du harcèlement est expliquée, dans le niveau individuel, par la qualité du lien entre l'élève et l'école et par l'adhérence à des comportements déviés; au niveau scolaire, par des caractéristiques contextuelles, telles que le climat général de satisfaction, la qualité du rapport enseignant-élève et la méthode de contention de conflits.

Mots-clés: Harcèlement; Écoles; Enseignement Primaire; Enseignement Secondaire

Resumen

El artículo examina el fenómeno del bullying entre los alumnos de la educación básica de las escuelas provinciales de Minas Gerais. El análisis empírico se basó en datos que se obtuvieron junto a 5.300 alumnos y a 243 profesores durante 2012 y 2013. Los datos se analizaron en base a modelos lineales generalizados jerárquicos. Se estimó la posibilidad de ocurrencia de la práctica de bullying a partir de una estructura de dos niveles, el de los alumnos y el de las escuelas. Se verificó que la distribución del bullying se explica, a nivel individual, por la calidad del vínculo del alumno con la escuela y la adherencia a comportamientos desviantes; a nivel escolar, por características contextuales, como un clima general de satisfacción, calidad de la relación profesor-alumno y método de contención de conflictos.

Palabras Clave: Bullying; Escuelas; Enseñanza Primaria; Enseñanza Secundaria

Bullying is characterized by Olweus (1997) as the repeated exposure of a student overtime to negative practices from one or more classmates. The negative practices the author refers to include all action applied by someone, intentionally or unintentionally, that causes damage to, hurts or bothers another person. It may be manifest through words (threats, scorn, offensive nicknames), physical contact (beating, pushing, slapping, hair pulling, pinching, obstructing passage of another) or through more subjective and insinuating ways (obscene gestures, exclusion, refusal to fulfill the other's desires).

This article focuses on bullying practices among students in basic education in the public school system of the Belo Horizonte Metropolitan Region - MRBH - and hub cities in the state of Minas Gerais. The empirical analysis relies on data from 5,300 students and 243 teachers in 87 public schools in Minas Gerais. This data results from the study Violência nas escolas [Violence in Schools], conducted by the Centro de Estudos de Criminalidade e Segurança Pública of Universidade Federal de Minas Gerais [Center for Studies on Criminality and Public Safety at the Federal University of Minas Gerais] - Crisp/UFMG -, between 2012 and 2013.

Due to its universal character, which is recurrent and harmful to physical, psychological and social health of students and, consequently, to the good functioning of educational activities, bullying among students has been motivating systematic studies since the 1970s. During the most recent decades, academic production on the subject has shed light on the universality and the expansion of bullying. Studies done in different parts of the world reveal victimization rates in elementary, middle and high school that vary from 4.2% to 49.8% of students. The aggression rates for the same public vary from 3.4% to 49.7%. The large divergence between values is due to the use of different methodologies and operational definitions, among other factors. Yet, the fact is that the problem has been detected, in larger or smaller measure, in every part of the world in which bullying has been studied (DAKE; PRICE; TELLJOHAN, 2003).

Considering the universality and harmful character of bullying practices, this work seeks to build an exploratory analysis of the phenomenon among students in their last years of elementary and high school in Minas Gerais' public schools, observing the prevalence of the problem, its possible causes, the profile of the students involved with it and the school dynamics and characteristics that influence bullying.

The intention is to develop a broad approach to the problem by incorporating contributions from fields of study and analysis of correlated phenomena. The work starts with a review of broader theoretical perspectives that, on one hand, shed light on the topic of school violence and the complexity of relationships built in this environment and, on another hand, on the deviant behavior among young people in general, and is followed by a focus on the problem of bullying itself.

Bullying: concept and academic production

The term bullying13 means "rowdy, troublesome, tyrannical". The first systematic academic studies on the phenomenon took place in Scandinavian countries around the 1970s. Since then, those countries have carried out longitudinal studies with the goal of understanding the effects of bullying (CATINI, 2004).

Olweus was the pioneer of the field of systematic studies on bullying and those facing it. In the 1980s, he carried out ambitious research, involving about 84,000 students, 400 teachers and 1,000 parents, seeking information on the scope of bullying among students, its characteristics and interventions made. The author found that one in every seven students was involved with bullying. A program to combat the problem was then developed that was capable of reducing bullying episodes by 50% in Norwegian schools (ZOEGA; ROSIM, 2009). This experience attracted attention from other European nations, such as the United Kingdom, Portugal and Spain, which have developed successful actions for prevention and handling such behavior during the 1990s. In 2001, the subject was a topic of interest of the European Economic Community, which implemented a bullying prevention project in several countries (ZOEGA; ROSIM, 2009).

As the subject has only received greater attention from researchers in the last decades, the definitions used by different authors frequently diverge, especially in terms of classifying bullying. There are different concepts: some divide it into only two categories - direct and indirect bullying or physical and non-physical, with others using three categories - physical, verbal and social bullying (also known as relational or indirect). Yet, the most-used definition among European and North American researchers is the one presented by Olweus (1997, p. 496), who says that "a student is being bullied or victimized when he or she is exposed, repeatedly and over time, to negative actions on the part of one or more students". Malta et al. (2010, p. 3.066) offer a very precise definition of the term, as a set of:

[...] behaviors of different levels of violence that go from inappropriate or hostile bothers to frankly aggressive acts, in verbal or non- -verbal form, intentional and repeated, without apparent reason, provoked by one or more students toward others, causing pain, anguish, exclusion, humiliation and discrimination, for example.

Marriel et al. (2006, p. 37) characterize it as "repetitive acts of oppression, tyranny, aggression and domination of persons or groups over other persons or groups, subjugated by force of the first". Therefore, the levels of violence within bullying behaviors range from teasing to demonstrations of hostility to verbal or physical aggressions with strong offensive potential.

The academic production on the subject frequently classifies bullying into three categories: (1) physical, which involves bodily offenses from the lightest (pushing, slapping), to the most serious (aggressions with guns); (2) verbal, which implies different verbal offenses with the intent of humiliating and intimidating the victim and; (3) social, which constitutes stigmatizing practices (gossip, lies, nicknames) and practices of exclusion. These are the categories used in our study. Although bullying practices are not restricted to the school environment, the focus of this article is the occurrence of bullying among students in elementary school to high school.

Bullying among students has become an object of attention among different researchers over the last few years. Different researchers (PHILLIPS, 2003; ZALUAR; LEAL, 2001; MALTA et al., 2010) show considerable association between involvement with bullying and negative consequences to physical, psychological and social aspects of the students and to the students' school performance (DAKE; PRICE; TELLJOHAN, 2003).

The most commonly attributed consequences for the psychological or social context of the victims are decreased or loss of self-esteem, an increase in feeling unsafe, and elevation of anxiety and depression. Additionally, the feeling of insecurity that it creates tends to decrease interest in education and motivation to attend class, therefore compromising school performance, learning and school attendance rates, and subsequently causing truancy, nervousness, difficulties in concentrating and even the possibility of self-mutilation and suicidal tendencies.

However, the effects do not just involve the students themselves. The teacher is another actor who can be deeply affected by constant episodes of bullying in the school environment. With bullying, the school becomes a violent environment where there is a lack of respect, ethics, social cohesion and solidarity. The impotence of the teacher and of the students in the face of attacks and a general perception of violence therefore compromises the educational process as a whole.

The school institution as a social agent

Discussing violence in the school context implies seeing the school as a social space, a place of social interaction and creation of ethos. It is necessary to observe the dynamics of daily interactions in the school, overcoming the notion that this institution is only a place for theoretical learning of subjects that compose the school curriculum. The school environment is not restricted to learning of content, but is also an environment of appropriation and development of meanings and interpretations of the world and of life itself. Students and teachers are sociocultural beings, involved in daily interactions that determine, in large part, the attitudes within institutions, norms and the contents that are transmitted.

Every individual defines his or her identity from the relationships established with others. The idea of "me" always refers to the notion of "others". There is no sense of self that does not directly refer to others. What is understood as "me" is not something isolated from the collective and social world, but the part that is of the greatest interest to the individual, precisely because this subject is both individual and general at the same time, in the intersection between the individual and the whole (COOLEY, 1992).

In this way, throughout the childhood and teenage years, the relationships established inside the school are fundamental to definition of the idea of "me", the image that the student has of his or herself, as well as opinions, aspirations and attitude within society. Because it is part of the social context and it is experienced from within, the school produces and redefines patterns of identity and behavior that are aligned or unaligned with what is deemed socially accepted and publicly defended as morally correct practices. These patterns can deeply influence the construction of the student's identity and personality and attitudes for the rest of his or her life.

The school is the institution where citizens learn to relate to others (GVIRTZ; BEECH, 2009). Prior to entering school, the individual receives influences nearly exclusively from family and the neighborhood, since interaction is basically restricted to those spheres. Upon entering school, the child's social interactions expand considerably. In school, the child has the opportunity to interact with classmates from totally different family settings, other social classes, other religions, etc.

In this environment, students are led to interact with those who are different and to learn to relate in society as citizens. This makes school a privileged locus of promotion of social cohesion. Gvirtz and Beech (2009) propose thinking of the school as a microcosm of society, in which the future citizens learn to relate in teams and with norms. This perspective sheds light on the matter of what type of social cohesion and sociability schools are promoting through their curriculums and institutional arrangements.

In a study conducted based on data analysis from the Program for International Student Assessment - PISA -14 the authors observe a high degree school enrollment segregation via socioeconomic as well as racial factors in all of Latin America, which means that Latin American schools are not achieving significant interaction between children and young people of different social classes. The conclusion is that the "Latin American educational systems seem to be collaborating in social fragmentation more than solving it" (GVIRTZ; BEECH, 2009, p. 357). Therefore, despite potential to create social cohesion among students and in society as a whole, the Latin American school systems need to progress a great deal on this matter.

The school institution and bullying

In this paper, we consider bullying through two aspects: individual characteristics of the student, especially those who exhibit violent and aggressive behavior, and contextual characteristics of schools, which, even if indirectly, encourage violent contexts.

Although most of the authors who study the practices of bullying among students focus on individual aspects or particularities of interpersonal relationships of the teenagers, reflecting on stigmas present in society as a whole, our interest lies primarily in the school as a social institution with specific characteristics and contexts that permeate the relationships among students and, which therefore, exerts influence over manifestations of violence and intimidation occurring among all actors involved in the institution.

Therefore, the discussion about the school as a complex institution, which has specific cultural aspects, organizational climate, hierarchical structures and ways of treatment that vary from one unit to the next, interests us as it sheds light on the school-setting aspects that can produce contexts that are more or less favorable to the practices of bullying.

Bullying is not interpreted in this work as an isolated relationship of domination between an aggressor and a victim, but as a manifestation of general tensions in a specific school context. The identification of many students who practice and suffer bullying in a specific school indicates the existence of a general climate of tension, which is not restricted to the linear relationship between aggressor and victim, but says something about the totality of interpersonal relationships built there.

Above all, the presence of bullying indicates the existence of structures of informal domination, in which some students subjugate others, exerting a violence that deviates from the school's objective of being an institution focused on society, learning and formation of citizens. Bullying also indicates the incapacity of the institution to promote an environment of healthy interaction among students. Certainly, a school with high rates of bullying has organizational and interrelationship problems, bullying being just one of the visible aspects.

The bibliography causes us to reflect on the organizational aspects of the school that may create a context more or less favorable to bullying. Flores-González and Retamal-Salazar (2011) highlight that a positive school climate, characterized by a general feeling of satisfaction and belonging, is inversely connected to the contexts of violence and school victimization.

Gvirtz and Beech (2009) call attention to the fact that specific organizational arrangements - curriculum, programmatic contents, interpersonal treatment, and hierarchical structures, among others - promote contexts of sociability more or less focused on cohesion and social well-being. We also consider how schools deal with the students and the contexts of violence that are established between them. The teachers and principals have decision-making power for specific situations that give them the ability to create a context of social cohesion and justice or one of fragmentation, discrimination and hostility. Since the presence of bullying indicates interruption of social harmony among the school's students, we aim to discover how the teachers' attitude and treatment of students, within the daily routines and for specific cases of intimidation, affect the magnitude of this phenomenon.

Do schools with high incidence of bullying take responsibility for intervening in specific cases of aggression and intimidation and for restraining violent practices among their students or do they outsource that responsibility to institutions specifically aimed at maintaining law and order? Silva and Salles (2010) highlight that criminalization of students' acts generates stigmatization, removes dialogue from the educational process, harms the construction of citizenship and generates feelings of hostility among school representatives of and student families. In light of this discussion and the analysis of empirical data, we aim to observe this and all other aspects indicated above and their relationship to bullying.

The theories of criminality /delinquency among young people and bullying

Although bullying does not appears as a crime in the Brazilian Criminal Code, this phenomenon can be interpreted in an way that is analogous to one of the theories of criminality and juvenile delinquency, since bullying also represent forms of conflict, imposition of the will of one individual over another, and thereby always implying the intention of hurting the other and possibly presenting different levels of offensive potential.

Delinquency is defined by Gottfredson (2001, p. 4) as "behavior that involves the use of strength or fraud, acts of defiance and disobedience, and acts that deliberately cause harm to the person or to others". In this paper, we have identified three theoretical approaches for explanation of this phenomenon, which may help us comprehend the practices of bullying, which are: the theory of self-control from Gottfredson and Hirschi (1990); the theory of social control, from Hirschi (1969) and the general theory of tension, from Agnew (1992). These theories have been widely employed in the study of the bullying phenomenon.

The three theoretical constructions present interesting and measurable elements that explain this deviant behavior in general and that, in this paper, are applied for investigation of the characteristics of students in Minas Gerais public schools bully their classmates. In characterizing the individual with low self-control, Gottfredson and Hischi (1990) offer a series of behavioral characteristics that may identify them, and which can be found in bullies. The theory of low self-control presents as a characteristic weak parental supervision during the childhood and teenage years. In this paper, we have analyzed the effect of parental supervision of their children's school life in terms of probability of the child's involvement in bullying. Regarding aspects of the young person's behavior that refer specifically to school,15 the theory points to lack of commitment and interest in school as indicators of low self-control and, therefore, the tendency towards deviant behavior. In this paper, we have also investigated the effects of those aspects on the practice of bullying. The association of groups of deviants is also a factor highlighted by the authors, which can be investigated based on identification of whether or not the student who practices bullying belongs to gangs or not.

The theory of social control, from Travis Hirschi (1969), interprets deviant behavior not from the point of view of self-control, but from the perspective of an individual's bond with moral expectations of society, which forces people to behave in a morally correct way. We would consider, based on this theoretical proposal, that involvement with bullying may be interpreted as a consequence of the young person's weakening bonds of social control, since those who bully show indifference and detachment in terms of moral values, as they subject and suppress the right to well-being of the victim.

Hirschi (1969) highlights the role of the school as a social institution that creates in the adolescent a feeling of conformance to the moral values of society and to the social roles that the school attributes to him or her and expects the individual to undertake. In this way, the non-conformity of the teenager with the school institution itself, which is manifest through lack of affection for the school and its representatives, among other aspects, would be an important indicator of the tendency towards deviant behavior. With the goal of evaluating this aspect, we have analyzed the effects of the level of conformity of the student to the school institution in terms of probability of involvement in the practice of bullying.

An important element highlighted by the social control theory, as well as by the self-control theory, is the strong association observed among different types of deviant behavior; in other words, a connection is easily found among different types of deviant behavior of a certain individual.

The general theory of tension, of Agnew (1992), considers deviation as the result of a situation or long-lasting context of tension, usually caused by feelings of rage and rebellion, from relationships that bring negative notions or stimuli to the young person. The tension may be created by failures in reaching goals or lack of positively valued assets, by suppression of positive stimuli toward the young person and by the introduction of negative stimuli.

Among the negative stimuli that can lead to deviation, Agnew (1992) mentions some related to the experience of the young person in the school environment, such as: derogatory treatment from teachers or classmates, verbal insults, physical aggressions, inappropriate or negligent treatment from teachers, excessive or humiliating punishments, among others. In terms of generating contexts of tension in the school environment, these stimuli may be associated with involvement of the students in bullying, reason for which they are analyzed in this paper.

Data and methodology

This article provides an exploratory analysis of the bullying phenomenon among basic education students in public schools in Minas Gerais. The empirical analysis developed relies on data obtained from a self-applied and completely anonymous questionnaire. As stated by Phillips (2003, p.714), this method is more effective means to obtain relevant information on deviant behavior from a respondent.

The sample considers schools as units of observation. The initial population is formed by all public schools that teach that second cycle of elementary school (roughly corresponding to "middle school" from 6th to 9th grade) and/or high school, from 10th to 12th grade in the Municipal System of Belo Horizonte and hub cities16 in the state of Minas Gerais.

Based on this population, in each school the classes that would respond were chosen randomly. In each selected class, the total of students and teachers who were present received a self-applied questionnaire to be responded to within the timeframe of one class (50 minutes). In total, 87 schools were visited in every region of the state, creating a database with information on 5,300 students and 243 teachers in public schools of Minas Gerais.

To verify the associations among the variables, the Hierarchical Logistic Regression Model was used, which allowed the creation on inferential analysis, which considered not only the individual characteristics that could be associated to the practices of bullying, but also the school characteristics that would explain the variants of the phenomenon in schools.

Variable answer: practice of bullying

Below we present information on the construction of the main variable of interest of this study: the indicator that allows for identification of students who practice bullying. The questions included in the questionnaire were formulated in order to diagnose episodes of three classifications of bullying presented in the international literature - physical bullying, verbal bullying and social bullying.

The question focused on identifying what exactly the practice of physical bullying was: "In the last month, have you attacked someone by pushing, slapping, throwing objects at the person or breaking this person's objects in this school?" To measure verbal bullying, the following question was presented: "In the last month, have you humiliated, offended or intimidated someone in this school?" In its turn, social bullying was measured through the question: "In the last month, have you excluded someone during group activities, called someone by offensive nicknames or made up lies about someone in this school?" In all three cases, the students who marked the option "Yes" were asked: "How many times has this happened?" The goal of this question was to identify the cases in which the aggression happened occasionally, only once, and to disregard them as indicators of bullying. Since the phenomenon is defined by repetition of determined behaviors, only cases that featured repeated behavior was classified as bullying.

The general bullying indicator was built by adding the three mentioned variables after weighting each one of them by the quantity of occurrences practiced by the student. The indicator therefore considers value 1 for the students who have answered "Yes" to at least one of the questions referring to the practice of bullying and have indicated two or more in response to the question "How many times has this happened?".

Results and discussion

In this section, the results of the empirical analysis developed in the study are presented. Table 1 demonstrates the proportion of students who practice each type of bullying, as well as the distribution of indicators of practices of bullying and victimization "in general", in other words, unifying the three variables into one indicator.

The indicators of bullying practice show that 9.6% of the interviewed students practice bullying, among which 5.3% practice physical bullying, 5% practice verbal bullying and 3.7% practice social bullying. Although these percentages seem low,17 they are consistent with previous findings in Brazil and around the world (OLWEUS, 1998; FANTE, 2005; ZOEGA; ROSIM, 2009).

Table 1:Distribution of the variables of bullying practice 

Distribution of the variables included in the statistical models

Although 5,300 students distributed in 87 schools were interviewed, after the specification of all for level 1 and 2 variables and the exclusion of the unanswered cases with invalid answers for those variables, the database was reduced to 3,252 students distributed in 79 schools. The number of schools included in the analysis was also reduced, since not all of them had the minimum number of interviews to adapt to the hierarchical logistical model, with all schools with less than 10 interviews excluded from the database. Below, we present the statistical descriptions of the variables included in the models.

Table 2: Distribution of qualitative variables 

Table 3: Distribution of quantitative variables 

Analysis of the models' statistical results

The null adjusted model allows the comparison between the components of the variable in Y which refer to the variable in level 1 (between students) and in level 2 (between schools).

Since the variation of the probability obtained for level 1 was σ2=0.90597 and for level 2 τ00=0.26950, we found that approximately 22.93% of total variation in the probability of practicing bullying is due to the characteristics of the schools. From the statistical significance of τ00, we have also concluded that all of the schools have different probability averages for the dependent variable. These results confirm the existence of variability among groups, therefore meaning that there is an empirical justification for the application of the hierarchical model.

Interpretation of individual level indicators

The models in which only the characteristics of individual levels were included presented interesting results that confirm a major part of the considerations from the bibliography and shed light on the discussion on other matters that are not completely established in the field of study. The statistical significance obtained from the group variables referring to deviations highlight the appropriateness of criminality and delinquency theories mentioned in this paper for the interpretation of bullying behaviors among students. The indicator of student conformity to the school environment also presented the expected result, especially in regards to the theory of social control.

Table 4: Results from models with level 1 variables 

The group's coefficient variables for sociodemographic characteristics confirm in part the information contained in the literature on bullying and on school violence in general, especially in what refers to the behavior by indicators of sex and education level. The models point out that male students in elementary school - as opposed to those in high school - have a higher chance of practicing bullying. Olweus (1998) found the same results. In the sample analyzed by the author, the participation of boys was more frequent considering any type of bullying and, in general, students in elementary school practiced this kind of aggression much more frequently than those in high school.

The variables for social class present an interesting result, which sheds light on a still open-ended discussion in the studies about bullying, involving the effect of the student's social status on the probability of involvement with bullying. The estimated coefficients in models for socioeconomic level point out that students with higher socioeconomic levels have significant higher chances of practicing bullying.

The variables indicating the mothers' educational levels also present statistical significance, in the sense those students whose mothers have higher education levels practice more bullying. The students who did not know the answer regarding their mothers' educational levels18 or who have answered that their mothers have not completed high school education have lower chances of involvement with bullying as aggressors, in comparison to those whose mothers have completed high school or have a higher education degree. The variable for students whose mothers range from completion of 9th grade and some high school did not present statistical significance. Therefore, the chance of bullying among these students does not present considerable divergence in relation to what was found for students whose mothers have the highest education levels.

In a general manner, both in terms of socioeconomic level and mothers' educational level, the data indicates that bullying practices tend to reproduce status asymmetries and hierarchies present in society. In other words, there is a consistent positive association between the educational level of the mother and the student's socio-economic status and the probability of the adolescent practicing this kind of aggression.

If we consider that the school environment reproduces cultural and structural aspects of the society to which it belongs, we can affirm that bullying, in a certain way, reproduces the hierarchical and symbolic structures of domination in Brazilian society in general. We can assume, for example, that students of higher socioeconomic levels practice more bullying because they feel superior to their classmates, or because, as exemplified in Bourdieu's (1998) positioning on symbolic school violence structures, feel more comfortable in the school environment and more free to intimidate students who do not belong to the same level. Although we do not have direct evidence, this result is consistent with the well-known conceptions that school frequently benefits students of better social status compared to those in a more unfavorable socioeconomic situation.

The final variable included in the student socio-demographic characteristic group, the family setting, presented negative association to bullying practice at first, but lost statistical significance when other variables with higher explicative power were included. We have therefore found that a student belonging to a two-parent, one parent or any other family setting does not create any effect on chances of involvement with bullying, according to our models.

We now focus on the interpretation of the variables' co-efficients on the relationship of the student and relatives with the school. The factor that measures parents' attention to a student's school life did not present any significance in the models in which it was included. The failure of this variable, as well as the one on family setting, in explaining the students' involvement with bullying, may indicate that the persistence of many teachers in blaming families for teenagers' bad behavior in school is disconnect from reality, especially because of school influence or interactive nature, as the mentioned status inequalities. The idea of a "dysfunctional family" as the main reason for the student's undisciplined and violent behavior does not find support in this study. As we will read in the pages ahead, in the interpretation of coefficients for the variable on conformity with the school environment and other level 2 variables, the quality of the relationship with school environment itself, and not within the family, provides better explanations for practices of bullying among students.

The socialization variable also did not present statistical significance, which indicates that even students who like and get along well with most of their classmates may practice bullying toward some specific students. The factor that measures the conformity of the student with the school environment - affection towards the school in general, towards the principal and toward the teachers - obtained negative association to practices of bullying, indicating that students who show less conformity have higher chances of practicing bullying. This result is consistent with Hirschi's (1969) approach on social control, since it interprets the involvement of teenagers with deviant behaviors as a consequence of weakened bonds of society's moral expectations and that force students to act in a morally correct manner.

The students who showed low affection towards the school and its representatives will probably not be that concerned with matching their behaviors to the expectations of these representatives and, therefore, will feel more comfortable in behaving in an undisciplined and violent manner. The approach of self-control also highlights the importance of the school as one of the institutions that most actively cultivates self-control (and consequentially contains deviant behavior) in children and teenagers, because of the institution has the opportunity to monitor and shape the students' behavior on a daily basis. Since the bond between the student and the institution is weakened, the assumption is that there is no longer exercise of self-control, which certainly increases the chances of involvement in deviant behaviors and, in our case, in practices of bullying.19

The final group of individual-level variables includes deviant practices: drug use, participation in gangs and presence of distorted values. The estimated coefficients show that all of these characteristics strongly associate with the practice of bullying. This result may be interpreted again in light of theories on self-control and social control, which highlight that the involvement of young people with "less serious" non-conforming behaviors are important predictors of the student's involvement in more disturbing deviant behavior with higher offensive potential.

For the self-control theory, these behaviors demonstrate student's low self-control. As for the social control theory, they demonstrate nonconformity to conventional social control. However, whatever the most appropriate analytical explanation may be (and this definition would require further research), the implications of this result are clear in terms of administration and formulation of public policies that are more focused on restraint of bullying practices. The main conclusion we gathered based on an analysis of the variables is that the student's involvement with bullying is associated with demonstrations of disregard of morally conformist behavior in general. This finding highlights the importance of the school paying attention to such students, monitoring their behavior in effective ways, seeking ways of guiding them towards non-deviant, non-violent behavior and of improving their relationships with the school itself.

Interpretation of school level indicators

As explained, after estimating level 1 variable models only, we selected the most appropriate model, eliminating the variables whose coefficients were not statistically significant at least 90% reliable, looking to obtain a more complete, but careful model. This model is the final model, of which the results are shown in the last column of Table 4. Based on it, we started the insertion of the level 2 variables. Therefore, the coefficients for the schooling level variables presented later, in Table 5, will be analyzed taking into consideration that variables are being controlled in level 1 by the presence of indicators of the final model.

Table 5: Results from models with level 2 variables 

As expected, considering that the greatest (77.07%) divergence in our dependent variable is explained by individual-level characteristics, a lower number of variables for school level found statistical significance in the estimated models. Yet, some important indicators were found to be significant as expected, confirming some affirmations in the literature about school climate, school violence and bullying.

In the group of variables that unite the institutional characteristics of schools, the presence of police and attending the morning school shift (in comparison to the night shift) showed a positive association with higher levels of bullying. The afternoon shift did not present significant divergence in comparison to the night shift, nor to the morning shift.20 Therefore, we conducted a correlation test between the four variables to verify if the afternoon shift, in fact, lacked any association to the practice of bullying or if the statistical significance was being affected by the inclusion of the morning shift. It was found that, while the morning shift has a positive correlation to the presence of bullying and the nightshift has a negative correlation (both statistically significant at 95% reliability), the afternoon shift does not show any association to the variation of the bullying factor.

We have yet to interpret the divergence between the morning and night shifts as it refers to the distribution of bullying practices. This difference is probably justified by the fact that the night shift involves students in higher grade levels - almost always in high school - therefore involving more mature young people. Also, a great part of those who attend the night school shift works during the day, while those who attend the morning shift are almost always dedicated solely to school. This setting may represent significant differences in maturity levels of the young people who are in one or another shift, being that, when entering the job market, the young person gains responsibilities that increase maturity, possibly affecting the way they relate to classmates at school.

The variable that measures the presence of extra-curricular activities was not found to be significant, which was not expected, considering that there is a general expectation that the offering of educational and leisure activities out of school hours improves the school's climate in general and helps contain violent contexts. A hypothesis for the non-significance of the indicator in our study is the fact that it has a positive correlation of0.313 statistical significance for the shift factor (morning), which may generate co-linearity between the indicators. However, a more careful analysis on the relation between offering of extra-curricular activities and the presence of bullying in schools must be the objective of future research efforts.

The presence of police in the school has presented a positive association with existence of bullying. Although it is very rare for the school to call on police to solve bullying problems between students - as shown in the distribution of variables of bullying repression presented in the section "Distribution of the variables included in the statistical models" -, we see later on that schools with high levels of bullying are also those with greater occurrences of other types of violence, such as physical aggression, theft, robberies and possession of guns. The presence of the police in these schools may be more a function of these other deviant behaviors than due to practices of bullying.

Still, the presence of police may indicate a distance in general between school representatives and their students. As noted by Silva and Salles (2010), criminal acts by students represent the exclusion of dialogue in the educational process, possibly even meaning prejudice and hostile feelings of teachers and principals towards the students, which harms the construction of a healthy socialization climate in the broader school environment.

The variables related to teacher behavior, in general and specifically in relation to bullying, found results that confirm the hypothesis that the better means to combat the problem are not repression and punishment, but construction of healthy personal interrelations based on dialogue, and not aggressiveness.

The aggressive teacher factor, for example, has presented strong association with the highest presence of bullying. The punitive teacher variable, although losing significance after the inclusion of school context variables, also presented positive association with bullying in models 2 and 3. As pointed out by Agnew (1992) in the tension theory, bad treatment by teachers and verbal insults directed at the students generates permanent contexts of tension and rebellion in the school environment, which may lead to student involvement in violent behaviors such as bullying, which serves an outlet for this tension.

Hurford et al. (2010) showed that school contexts in which the teacher-student relationship is based on inequality and low teacher receptiveness and interest in students involves higher levels of bullying, feelings of insecurity and threats with guns. Stoll and Fink (1999) also pointed out the importance of prioritizing a teacher-student relationship based on dialogue, inclusion and respect, to create a school environment that is not only peaceful, but also effective in its mission of socialization and content transmission. This new type of relationship must substitute the one based on strict authority structures, aggressiveness and intimidation.

The variables for addressing bullying in school do not obtain statistical significance in the models nor did they present high levels of correlation with other inserted variables, which discards the hypothesis of co-linearity. When faced with the complexity of the inter-relationships experienced daily in school, the simple fact of addressing the subject loses importance. It is possible that combating the problem is much more due to change of posture of the actors in daily practice than through the simple inclusion of the subject in the contents to be covered in class. As well as repressing bullying cases through strict measures, such as suspension or calling on distant authority structures, do not seem to have any success - as shown by the non-significance of the variable of severity of the repression -, simply addressing the subject in a general way in the classroom also seems to be a weak measure.

The coefficients of the variables in the last group, referring to the school context, point out that the solution involves broad approaches, which seek to change the climate and the culture of the school overall. The violence factor, which measures the existence of a general context of deviant behavior in the school, obtained a strong positive association with the presence of bullying. The school climate factor, in its turn, obtained negative association with bullying, which means the higher the sense of satisfaction of the students toward the school environment, the fewer the chances of bullying practices in these schools.

Flores-González and Retamal-Salazar (2011) show that violence in school environments is influenced by the dynamics of the experiences and relationships in those surroundings overall, so that the school climate influences conflict or oppressive situations among students. The authors also point out that the quality of the school climate, measured as the feeling of belonging, pride and satisfaction of the students towards the school, is inversely related to the victimization context, so that maintaining a positive school climate is a way to prevent or reduce school violence, being that a negative school climate predicts violence in that environment.

The interpretation of the variables that obtained significance in level 2 lead to the conclusion that the school contexts favorable to bullying are those that are also contaminated by the existence of other types of negative influences, such as unfavorable teacher attitudes, general feeling of dissatisfaction, and dissemination of different deviant practices. The results support the approaches that propose the combating of bullying through broad initiatives that seek to improve the school environment as a whole and all of its complex net of interactions.

Finally, we present in Table 6 the results of the consolidated model, which unites only the variables that presented statistical significance in the previous models. Since the relationships highlighted by this model have already been interpreted in the previous paragraphs and the coefficients have not undergone any alteration of meaning or significance in comparison to the others, we will not dedicate ourselves here to a new interpretation of the results. We will only introduce one measure of adjustment of the model to find if the group of considered characteristics explains the distribution of probabilities of bullying in the schools in a satisfying way.

We have calculated the proportion of variance explained by β0j of the model with the inclusion of level 2 variables, comparing the variance between schools after specifications of this model, with the variance found in the Null model. The proportion of variance is determined by:

It measures how much the β0j of level 2 of the consolidated model explains the total of variation of the parameter in the average of schools. It was found that 53.07% of true variance among schools for the probability of bullying practice is attributed to the explanatory variables of level 2.

Table 6: Results of the consolidated model 

Final considerations

This study focused on practices of bullying among public school students in Minas Gerais, with the objective of broad analysis of the issue. The greatest effort was made in terms of finding individual and contextual factors that could explain the distribution of bullying practices among this public.

Different revealing results were obtained from the estimates of hierarchical regression models, for individual student characteristics as well as school context. On the individual level, for example, we have found that the student's tendency to practice bullying is more influenced by the quality of the student's bond with the school and its representatives than by family characteristics, such as parental attention to the students' school life and the family setting itself. This is an interesting result as it breaks away from the concept defended by many educators that violent and undisciplined behavior of students in school is primarily a result of "dysfunctional families".

Another result that opposes this idea is the positive association found between socioeconomic level and the mother's educational level and the practices of bullying. The observation that young people with higher social status - considering the average status of students who go to public schools - practice more bullying, opposes the still-current notion that contexts of oppression and intimidation are more frequent among students of low socioeconomic levels.

This finding opens up a new research agenda, in that a more complete interpretation of the relationship between social status and bullying demands investigation of factors that we were unable to measure here, such as the social status of the victims and school dynamics that possibly generate a configuration of superiority among students of better socioeconomic levels over those less fortunate.

The school characteristics also confirm that the dynamics at the school itself are very important for the emerging of bullying behaviors. Factors like negative school climate -, understood as a general sense of students' dissatisfaction with schools, school representatives, and their own classmates -, aggressiveness from teachers, presence of the police and tendency to use punishment as a way to restrain conflict, showed a positive association with higher levels of bullying. These results confirmed the hypothesis that combating the problem should not be done via punishment and external interventions, but through construction of a healthy climate of personal inter-relationships based on dialogue, on humanized treatment and non-aggressiveness.

These results also support the conclusion of previous experiments and studies that anti-bullying programs that are universally focused, in that they work equally with all the students, are more effective in restraining the problem than those that focus only on the aggressors and the victims. Additionally, it becomes evident that anti-bullying initiatives must include the entire complexity range of dynamics and interactions in the school environment, as a way to reverse contexts of dissatisfaction, aggressiveness and violence as a whole: "School-based initiatives developed to reduce bullying behaviors should incorporate interventions designed to promote positive social interactions between students and teachers in particular, and between all members of the school community" (RICHARD; SCHNEIDER; MALLET, 2011, p. 278).

In sum, the paper shows that bullying behaviors are not an isolated phenomenon. On the contrary, the issue has been revealed to be a general one, despite variations perceived from one context to another within the entire state system. Additionally, it is a phenomenon that is not just formed by and responds to students' individual characteristics, inclinations or profiles. It is, above all, a component and a result of a group of social relationships established in the scope of the school institution and interactions that are formed there. Finally, the way in which the school responds and deals with the construction of an interactive climate and with occasional or systematic discipline may contribute to enhancing or mitigating this kind of practice. It is important to recall that the school is responsible not only for teaching academic content, but for the concept of interaction. The values transmitted by example and through the experiences the school itself creates are a fundamental part of the institution's educational mission.

Referências

AGNEW, R. Foundation for a general strain theory of crime and delinquency. Criminology, n. 30, p. 47-88, 1992. [ Links ]

BOURDIEU, P. Escritos de educação. Organização de M. A, Nogueira; A, Catani. Petrópolis, RJ: Vozes, 1998. [ Links ]

CATINI, N. Problematizando o "bullying" para a realidade brasileira. Tese (Doutorado em Psicologia) - Pontifícia Universidade Católica de Campinas, Campinas, 2004. [ Links ]

COOLEY, C. H. The social self. In: PONTELL, H. Social deviance: readings in theory and research. New Jersey: Prentice Hall, 2002. [ Links ]

DAKE, J. A.; PRICE, J. H.; TELLJOHAN, S. K. The nature and extent of bullying at school. Journal of School Health, v. 73, n. 5, p. 173-180, May 2003. [ Links ]

FANTE, C. Fenômeno bullying. Campinas, SP: Versus, 2005. [ Links ]

FLORES-GONZÁLEZ, L. M.; RETAMAL-SALAZAR, J. A. Clima escolar y gestión compleja del conocimiento: desafíos para la investigación educativa y la política pública en violencia escolar. Magis, Revista Internacional de Investigación en Educación, v. 4, n. 8, p. 319-338, 2011. [ Links ]

GOTTFREDSON, D. Schools and delinquency. Cambridge: Cambridge University Press, 2001. [ Links ]

GOTTFREDSON, M; HIRSCHI, T. A general theory of crime. Stanford: Stanford University Press, 1990. [ Links ]

GVIRTZ, S.; BEECH, J. Micropolítica escolar e coesão social na América Latina. In: SCHWARTZMAN, C.; COX, C. (Org.). Políticas educacionais e coesão social: uma agenda latino-americana. Rio de Janeiro: Elsevier; São Paulo: IFHC, 2009. [ Links ]

HIRSCHI, T. Causes of delinquency. Berkeley, CA: University of California Press, 1969. [ Links ]

HURFORD, D. P. et al. The role of school climate in school violence: a validity study of a web-based school violence survey. Journal of Educational Research & Policy Studies Spring, v. 10, n. 1, p. 51 -77, 2010. [ Links ]

INSTITUTO NACIONAL DE ESTUDOS E PESQUISAS EDUCACIONAIS ANÍSIO TEIXEIRA - INEP. Portal do Inep. Disponível em: <Disponível em: http://portal.inep.gov.br/pisa-programa-internacional-deavaliacao-de-alunos >. Acesso em: 00/02/2014. [ Links ]

MALTA, D. C et al. Bullying nas escolas brasileiras: resultados da Pesquisa Nacional de Saúde do Escolar (PeNSE), 2009. Revista Ciência & Saúde Coletiva, v. 15, supl. 2, p. 3065-3076, 2010. [ Links ]

MARRIEL, L. C. et al. Violência escolar e auto-estima de adolescentes. Cadernos de Pesquisa, São Paulo, v. 36, n. 127, p. 35-50, jan./abr. 2006. [ Links ]

OLWEUS, D. Bully/victim problems in school: facts and intervention. European Journal of Psychology of Education, v. 12, n. 4, p. 495-510, 1997. [ Links ]

OLWEUS, D. Conductas de acoso y amenaza entre escolares. Madrid: Morata, 1998. [ Links ]

PHILLIPS, C. Who's who in the pecking order? Agression and 'normal violence' in the lives of girls and boys. British Journal of Criminology , n. 43, p. 710-728, 2003. [ Links ]

RICHARD, J. F.; SCHNEIDER, B. H.; MALLET, P. Revisiting the whole-school approach to bullying: really looking at the whole school. School Psychology International, n. 33, p. 223-239, 2012. [ Links ]

SILVA, J. M. A. de P.; SALLES, L. M. F. A relação professor-aluno como um dos enfoques de análise da violência escolar. Educação em Foco, Juiz de Fora, v. 15, n. 1, p. 87-104, 2010. [ Links ]

STOLL, L.; FINK, D. Para cambiar nuestras escuelas: reunir la eficacia y la mejora. Barcelona: Octaedro, 1999. [ Links ]

UNIVERSIDADE FEDERAL DE MINAS GERAIS. Violência nas escolas e programas de prevenção: estudo dos possíveis impactos do programa "Escola Viva, Comunidade Ativa" nas escolas estaduais de Minas Gerais. Belo Horizonte: UFMG/Crisp, 2013. [ Links ]

ZALUAR, A.; LEAL, M. C. Violência extra e intramuros. Revista Brasileira de Ciências Sociais, São Paulo, v. 16, n. 45, p. 145-164, fev. 2001. [ Links ]

ZOEGA, M. T. S.; ROSIM, M. A. Violência nas escolas: o bullying como forma velada de violência. Unar, Araras, v. 3, n. 1, p. 13-19, 2009. [ Links ]

Received: April 2016; Accepted: July 2016

Creative Commons License Este é um artigo publicado em acesso aberto sob uma licença Creative Commons