Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

Violent relationships at the social-ecological level: A multi-mediation model to predict adolescent victimization by peers, bullying and depression in early and late adolescence

Violent relationships at the social-ecological level: A multi-mediation model to predict... Citation: Oriol X, Miranda R, Amutio A, Acosta HC, Mendoza MC, Torres-Vallejos J (2017) Violent relationships at the social-ecological level: A multi- Background mediation model to predict adolescent victimization From the social-ecological perspective, exposure to violence at the different developmental by peers, bullying and depression in early and late adolescence. PLoS ONE 12(3): e0174139. https:// levels is fundamental to explain the dynamics of violence and victimization in educational doi.org/10.1371/journal.pone.0174139 centers. The following study aims at analyzing how these relationships are produced in the Editor: Andrew R. Dalby, University of Peruvian context, where structural violence situations exist. Westminster, UNITED KINGDOM Received: October 30, 2016 Methods Accepted: March 3, 2017 A multi-mediation structural model with 21,416 Peruvian adolescents (M = 13.69; SD = 0.71) was conducted to determine the influence of violence in the school environment on Published: March 30, 2017 violence perceived within school and violence exercised by teachers. In addition, it was also Copyright:© 2017 Oriol et al. This is an open intended to determine whether these violent relationships predict depression through loneli- access article distributed under the terms of the Creative Commons Attribution License, which ness, and bullying through peer victimization. The existence of differences between early permits unrestricted use, distribution, and and late adolescence was also verified. reproduction in any medium, provided the original author and source are credited. Results Data Availability Statement: All relevant data are within the paper. Results confirm that violence in the school setting has high influence on violence exercised by adolescents and teachers within the school. Teacher violence is the most important pre- Funding: This research was made possible with the support of the Ministry of Education of Peru, dictor of depression through loneliness, and encourages peer victimization and the emer- especially the Management Division of Educational gence of aggressive behavior. Exposure to violence exercised by support sourcesÐteachers Quality (MINEDU) who were responsible for and classmatesÐexplains more than 90% of the total variance explained in bullying behav- carrying out the data collection within the ior. Differences were found between early and late adolescence models. framework of the impact assessment carried out by the Specialized System against School Violence (SiseVe). PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 1 / 15 Violent relationships: Bullying and depression in early adolescence Competing interests: The authors have declared Conclusion that no competing interests exist. The high prevalence of structural violence in school settings facilitates the bullying/victimiza- tion dynamics within school. From a social-ecological perspective, this result suggests the importance of network cooperation at a mesosystem level, with teachers from educational centers playing a crucial role in the prevention of bullying/victimization. Introduction From the social-ecological perspective [1], relationships between developmental environments and the dynamics that occur within them are fundamental to understand how the risk of vio- lence, its protective factors, and depression in adolescence are constructed (for revision, [2]). Thus, a central tenet of this theory is that individual development is influenced by the ongoing qualities of the child's social settings and the interactions between these settings (e.g., family, peers, schools, communities) [3]. An educational model oriented towards a conception of the ecological development of the individual cannot help but consider the varied forms of interac- tion and interchange that children experience both inside and outside schools [4, 5, 6]. In the social-ecological model, the mesosystem refers to the relationships between the dif- ferent microsystems. This concept is of paramount importance for studies on bullying [7, 8, 9] and, thus, for this specific study. Specifically, student-teacher relationships are microsystems that consist of the multiple interrelated perceptions that both parties have about their interac- tions [10, 11]. Perceptions are important because they are real, from a psychological stand- point, and they have the power to influence the behavior of each party significantly [12]. During adolescence, the school environment serves a pivotal role in the development pro- cess and asserts itself as one of the most important socialization spaces [13]. In addition, affec- tive relationships are established with peers and adults like teachers, who acquire special significance in this developmental stage [14, 15]. School, as a socializing setting, forms part of a neighborhood or geographical zone which, in turn, influences the experiences and everyday relationships at school [16,11]. Therefore, the study of the structural violence present in zones or neighborhoods where schools are located has become a growing interest over the last years, as it may help explaining bullying phenomena [17, 2, 16]. From the ecological perspective, understanding the influence of diverse factors on victimization/bullying dynamics as well as the relationships between the different actors becomes necessary for adopting effective preven- tive measures. The main factors to be studied are: 1) Contextual factors: According to the Theory of Structural Violence developed by Galtung [18], the extreme adversity conditions to which individuals are subjected within a society may be considered structural violence conditions when they are embedded in the social, political and economic organization of society. Peru is one of the countries where the Shining Path ter- rorist activity from 1980 provoked structural consequences, which makes it be regarded as a country ravaged by violence until today [19]. During decades, many of the population was fre- quently exposed to stress situations that generated important consequences such as depression, anxiety and post-traumatic stress disorders (PTSD) [20]. Currently, violence is considered a pervasive issue in Peru, which manifests in all contexts, such as workplaces, streets, public spaces, and affects equally men, women, adolescents and children [21]. Among Peruvian youngsters, these environments affected by structural violence translate into participation in street gangs and high consumption of alcohol from early ages. Furthermore, 50% of regular consumers start consuming alcohol at age 13, while 90% starts before 16, and this is PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 2 / 15 Violent relationships: Bullying and depression in early adolescence encompassed by a high prevalence of violence in schools [22, 23]. Schools located in settings where structural violence is entrenched might also reproduce certain violence dynamics within the school environment itself as pointed out by recent studies [24, 25]. 2) Relationships with significant support figures: Peer relationships that are mostly devel- oped during the time spent in educational centers come first in the list of adolescents' prior- ities [26, 27]. Adolescents need to be recognized by their peers, and the social support received from the latter is essential to strengthen their self-esteem [28]. Preadolescence and transition to secondary school are especially prone to bullying, due to the physical and psy- chological changes that occur during these periods [29, 2, 30]. Thus, when adolescents are perceived as different from others, whether because of their ethnicity, sex, socioeconomic status, the probabilities of being assaulted by their peers rise [31]. The same is true for psychological variables like depression [32], which may be one of the consequences of vic- timization, since violence, especially from peers, undermines self-esteem and also makes adolescents feel helpless and depressed [33, 34, 35, 32]. Additionally, psychosocial adjust- ment problems and low self-esteem sometimes make victims of abuse become aggressors [36, 37, 38]. Therefore, several studies consider victimization and bullying as part of a same phenomenon [39]. The consequences of bullying may be even more severe if there is a lack of socio-emotional support [40, 2, 41]. In this sense, the affective relationship that teachers establish with students are a key factor for school adaptation [42, 6, 43] and for other variables related to adolescent adjustment, such as social functioning (e.g., [44]), behavior problems (e.g., [45]), and academic achievement [46, 47, 48]. This is confirmed by a recent study conducted by Wang, Brinkworth and Eccles with 1400 youngster, showing that trust and positive affective relationships with teachers act as moderators in the relationship between parent-children conflicts, and depres- sion and bad behavior in 13-to-18-year-old adolescents [49]. In this sense, the type of interac- tions that adolescents establish with significant people, depending on their developmental stage, appears to act as important protective or risk factors for victimization/bullying [3, 9]. Present study The constant changes that adolescents undergone with the onset of puberty makes this a time of special vulnerability, which contributes to the emergence of bullying [2]. From a social-eco- logical standpoint, it is estimated that violence relationships at a mesosystem level may influ- ence the violence interactions that occur between peers during adolescence. In this regard, the structural violence of a context can turn affective relationships into risk factors rather than act- ing as protective factors [28]. Taking all this into consideration, the following hypotheses are proposed: (1) Perception of violence in the neighborhood where a school is located is expected to influence violence between peers within the educational center, as well as violence exercised by teachers in the classroom; (2) Violence relationships (i.e. between peers within the school and from teachers in the classroom) will be predictors of bullying behavior through victimization by peers; (3) Peer violence at school, violence exercised by teachers, and victimization will predict depres- sion through loneliness; and (4) Differences will be found between early and late adolescence in the aforementioned variables. Literature shows differences in aggressiveness and depression according to gender in ado- lescent population. Concretely, most studies report a higher prevalence of bullying cases in boys rather than in girls [50, 2]. Girls, on their part, would have greater levels of negative emo- tionality and depression [51, 52]. Therefore, gender was controlled for in the proposed struc- tural equation model. PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 3 / 15 Violent relationships: Bullying and depression in early adolescence Table 1. Demographic characteristics of participants (n = 21416). Characteristic n % Gender Male 11528 53.8% Female 9888 46.2% Stage of adolescence Early adolescence 14664 69.1% Late adolescence 6559 30.9% https://doi.org/10.1371/journal.pone.0174139.t001 Method Sample Participants were part of an impact assessment study on school violence developed by the Min- istry of Education of Peru, which sought to reduce the cases of violence in schools. Based on this criterion, Ministry of Education of Peru considered schools for this purpose and contacted directly with them'. In total, 21,416 13-to-17-year-old students in 70 schools from 19 different regions at a national level took part in the study (see Table 1). The mean age was 13.69 (SD = 0.71). As for the period of adolescence (early vs. late), most participants were in early adoles- cence. Thirty percent of the sample came from Lima province (i.e., Lima and Lima Callao), which has the greatest population of Peru (see Table 2). Procedure and ethics statement The Ministry of Education of Peru directly contacted with all educational institutions. Parents and principals were informed of the scope and importance of the study, which lasted four months. All questionnaires were administered in the classroom in the presence of at least one member of the research team. Table 2. Sample distributed by provinces. Province n % Amazonas 28 1% Ancash 931 4.3% Arequipa 604 2.8% Ayacucho 248 1.2% Cajamarca 102 05% Cusco 58 0.3% Huanuco 227 1.1% Ica 3.143 14.7% Junõn 2.655 12.4% La Libertad 1.799 8.4% Lambayeque 1.897 8.9% Lima 4.820 22.5% Lima-Callao 1.586 7.4% Pasco 233 1.1% Piura 376 1.8% Puno 804 3.8% San Martõn 570 2.7% Tacna 1.128 5.3% Tumbes 207 1% https://doi.org/10.1371/journal.pone.0174139.t002 PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 4 / 15 Violent relationships: Bullying and depression in early adolescence Ethical approval for this study was granted by the Ministry of Education of Peru (MINEDU). Parents were requested to sign an informed consent for the application of the questionnaire. To protect the privacy of the students, confidentiality of the questionnaires was guaranteed, and students consent was requested prior to their application. International ethical guidelines for studies with human subjects described in the Nuremberg Code and in the Decla- ration of Helsinki were applied. Measures Reason for being bullied. Based on questions about the reasons for being bullied by other classmates. This scale was extracted from the Single School Well-being Questionnaire-CUBE (in Spanish, Cuestionario Unico de Bienestar EscolarÐCUBE) [53, 54]. This questionnaire was elaborated during the impact assessment carried out by the Ministry of Education of Peru to assess the development of socioemotional skills in educational institutions from Lima Met- ropolitan Area. This scale has 10 items that explore the reason for experiencing bullying (e.g. socioeconomic status). Each question was dichotomous, with 0 accounting for (ªNo), and 1 indicating (ªYesº). Kuder-Richardson reliability was .99. Teacher's aggressive behavior. The Teacher Violence Scale was based on questions about victimization extracted also from CUBE [53, 54]. This scale is composed by five items that enquire the student on violence situations directly (e.g. A teacher threatened to hurt you or beat you) or indirectly involving the teacher as the aggressor (e.g. You saw a teacher insulting other student), and whose range of response goes from 0 (ªNeverº) to 2 (ªTwo or more timesº). CronbachÂsα for the studied sample was .90. Bullying behavior. The Bullying Behavior Scale was based on the CUBE scale. This scale was composed by four items aimed at assessing the experience of the student as an aggressor (e.g. I threatened to hurt or beat other student). The response range goes from 0 (ªNeverº) to 2 (ªTwo or more timesº). CronbachÂsα = .89. Victimization. To measure this construct two scales were used: (i) The School Well-being Questionnaire and (ii) the inventory developed by Espelage and Holt [55]. This eight-item scale assesses the frequency of victimization suffered by the student considering the following categories: physical, verbal, exclusion, and cyberbullying (e.g. One or more students insulted you). The response range is from 0 (ªNeverº) to 2 (ªTwo or more times) CronbachÂsα was = .92. Loneliness. Three-item was used to measure this construct, which is based on a brief ver- sion of the 20-item UCLA Loneliness Scale Test (e.g. You feel isolated from others). The responses for these questions range from 1 (ªrarely or neverº) to 4 (ªMost or all the timeº). The advantage of this shorter version is that its application is more practical and less expensive than that of the original scale. It is worth noticing that this version showed appropriate psycho- metric properties in previous studies [56, 57]. CronbachÂsα = .88. Violence in school environment. Three-item self-constructed scale was used to identify vulnerability to violence in the school surroundings. The items are: (1) There are places in the way to or back from school that I don’t like to go through because I’m afraid that somebody hurts me; (2) In my school many students are in gangs and; (3) Crime and violence in my neighborhood are affecting my school, and whose responses range from 0 (ªNoº) to 4 (ªalwaysº). CronbachÂs α = .71. School violence. Four items was used to assess this variable, based on the abbreviated ver- sion of the California School Climate and Safety Survey [58] (e.g. The students of my school get into fights), and with response categories ranging from 0 (ªNoº) to 4 (ªalwaysº). CronbachÂsα was .77 PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 5 / 15 Violent relationships: Bullying and depression in early adolescence Depression. 10-item Depression Symptoms Test developed by Bradley [59] was included. This test consists of a shorter version of the originally 20-item a CES-D [60]. An example item on the scale is: I felt that everything I did was an effort. Response options range from 1 (ªrarely or neverº) to 4 (ªMost or all the timeº). CronbachÂsα was .79. All scales included in this study achieved the cut-off point of .70 proposed by [61] Statistical analysis Descriptive statistics and correlations were calculated by means of the statistical program SPSS 22.0. The statistical package AMOS 18.0 was used for structural equation models, and multiple indirect effects were calculated using maximum likelihood estimation together with bias-cor- rected confidence interval bootstrap test. This procedure provides an average of the estimates obtained from bootstrap samples and their standard error. To verify the model fit, absolute 2 2 and relative goodness-of-fit indices of the model were used, that is:χ indicators and theχ /gl coefficient, Comparative Fit Index (CFI), Incremental Fit Index (IFI), Root Mean Square Error of Approximation (RMSEA) and Standardized Root Mean Square Residual (SRMR). A multi-group analysis was also run to verify differences between the early adolescence and late adolescence models. Results Descriptive analysis Prevalence of exposure to violence. First, percentages were obtained through the responses to the Likert-type scales. The prevalence of violence in school environment, violence into school, as well as the presence of violent gangs were calculated taking as reference the response option corresponding to ªsometimesº in the answers given to the scales measuring violence. In order to calculate the prevalence of violence suffered by teachers and peers, the answers that reported at least one aggression or abuse during the last month were considered. Results show that 9,209 students (43%) felt unsafe when walking around the school sur- roundings. In addition, 12,721 (59.4%) adolescents reported usually seeing violent behaviors in school and 10,494 (49%) informed that youngsters from school belonged to violent gangs. Furthermore, 8,673 (40.5%) students reported having suffered abuse or insults from their peers at school. Finally, 4,237 adolescents (20.3%) stated having suffered or observed some type of violence from the teacher in the classroom. Table 3 summarizes the reasons for bullying behaviors according to the responses of the dichotomous scale. When comparing gender, men presented higher levels of aggressiveness (M = .44 vs. M = .28; p < .001), whereas women present higher levels of depression (M = 2.90 vs. M = 2.00; p < .001). Correlations. Correlations were performed between all the study variables. As shown in Table 4, all correlations were significant and with values above .15. Measurement and structural indirect effects model Firstly, to verify the fit of the variables into the model, HarmanÂs single factor test was con- ducted [62]. Two models were evaluated: (1) One latent factor model, in which all items were included; and (2) multiple latent factors model, in which the variables considered in this study were included. The results showed an adequate fit to the second model proposed. The model fit is as follows: X²/df = 1.66, df = 199; p < .001; CFI = .94; TLI = .93; SRMR = .04; RMSEA = .04. Subsequently, a first structural equation model was conducted to test the proposed hypoth- eses of this study. In this model, violent relationships in the school were considered as predic- tors of bullying behavior through victimization, and victimization by peers was taken as a PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 6 / 15 Violent relationships: Bullying and depression in early adolescence Table 3. Prevalence of reasons for being bullied in male children (n = 11528) and female children (n = 9888). Male children Female children Targets n % N % χ2 a. Racial differences 975 8 717 7 6.17** b. Geographical differences 514 4 310 3 20.44*** c. Sexual condition 603 5 302 3 55.12*** d. Physical appearance 2434 21 1780 18 20.34*** e. Presence of disabilities 293 3 128 1 38.54*** f. Religious beliefs 806 7 538 5 16.14*** g. Socioeconomic status 1192 10 736 7 43.44*** h. School grades (for both high and low scores) 1872 16 1437 15 5.22* i. Without apparent reason 1917 17 1191 12 71.87*** j. Other reasons 2245 19 1734 18 6.32** * p< .05 ** p< .01 *** p< .001. https://doi.org/10.1371/journal.pone.0174139.t003 predictor of depression through loneliness as suggested in hypothesis two and three. Neverthe- less, it was observed that victimization by peers did not show a predictor effect on depression nor contribute to loneliness with more total variance explained. Therefore, a new model with- out that relationship was calculated. The results showed an adequate fit for the final model: X²/ df = 1.66, df = 199; p < .001; CFI = .94; TLI = .93; SRMR = .04; RMSEA = .04 (Fig 1) All direct effects of the model were significant (see Table 5). Regarding indirect effects with bias-correct confidence intervals, an indirect effect is observed in the relationship between school violence and depression through loneliness (CI = [.021, .134]); p = 0.01); and between school violence and bullying behavior through victimization (CI = [.029, .044]); p = 0.05). Fur- thermore, a significant indirect effect between teacher's aggressive behavior and bullying behavior through victimization (CI = [.350, .459]); p = 0.05), and between teacher's aggressive behavior and depression through loneliness (CI = [.013, .044]); p = 0.01) was identified. The variables included in the model explain 15% the variance of depression and 93% of the full model variance explained. Invariance across adolescence stage To calculate the invariance between the early and late adolescence models, a multi-group anal- ysis was conducted. Significant differences were identified between the unconstrained and Table 4. Correlations and CronbachÂsα between all the variables. Variable 1. 2. 3. 4. 5. 6. 7. 1. Violence in the environment Ð 2. School violence .36*** (.71) 3. Teacher's aggressive behavior .26*** .36*** (.90) 4. Loneliness .24*** .20*** .15*** (.88) 5.Victimization .29*** .38*** .81*** .17*** (.92) 6. Depression .21*** .19*** .17*** .31*** .18*** (.79) 7.Bullying behavior .26*** .35*** .91*** .15*** .86*** .17*** (.89) *** p< .001 https://doi.org/10.1371/journal.pone.0174139.t004 PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 7 / 15 Violent relationships: Bullying and depression in early adolescence Fig 1. Multi-mediation structural equation model (SEM). https://doi.org/10.1371/journal.pone.0174139.g001 measurement weight models (p < .001). Specifically, slightly stronger relationships in path estimates were observed in the early adolescence model, although total variance explained in both models remain similar (Table 6; Fig 2). In the two models, indirect effects through victim- ization and loneliness were significant. Discussion First, descriptive data show that almost half of Peruvian youngsters perceives school's environ- ment as unsafe. In particular, 43% feel unsafe when walking down the streets surrounding school, and 49% reports that students from their school belong to violent gangs. This data is consistent with a high prevalence of peer victimization. Specifically, more than 40% of the Table 5. Standardized direct effects of the model. Dependent Variables Independent Variables Direct path estimate School violence <Ð Violence in the school environment .84***[.823, .886] Teacher's aggressive behavior <Ð Violence in the school environment .31***[.291, .327] Loneliness <Ð School violence .41***[.378, .442] Depression <Ð School violence .09*[.065, .125] Victimization <Ð School violence .08* [.063, .090] Bullying behavior <Ð School violence .13**[.104, .150] Loneliness <Ð Teacher's aggressive behavior .07*[.063, .090] Victimization <Ð Teacher's aggressive behavior .87***[.854, .886] Bullying behavior <Ð Teacher's aggressive behavior .63***[.567, .690] Depression <Ð Teacher's aggressive behavior .17***[.128, .195] Depression <Ð Loneliness .28***[.254, .315] Bullying behavior <Ð Victimization .46***[.405, .528] * p< .05 ** p< .01 *** p< .001. Note: 95% intervals are in brackets https://doi.org/10.1371/journal.pone.0174139.t005 PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 8 / 15 Violent relationships: Bullying and depression in early adolescence Table 6. Invariance between models of early and late adolescence. 2 2 2 Models χ df χ /df Δχ Δdf CFI TLI SRMR RMSEA Model 1 14184 1162 - - .96 .96 .02 .02 Model 2 14305 1190 5.20 120.82* 28 .96 .96 .02 .02 Model 3 14353 1204 5.02 171.22* 43 .96 .96 .02 .02 Model 4 15470 1247 5.40 193.66* 50 .96 .96 .02 .02 * p< .001 Model 1: Unconstrained; Model 2: Measurement weights; Model 3: Structural covariances; Measurement residuals https://doi.org/10.1371/journal.pone.0174139.t006 sample informs having suffered from any type of physical or verbal violence, which widely exceeds data found by other studies reporting a prevalence of 20% (see review [2]). Likewise, almost 60% of students have observed any kind of violence in school, and more than 20% states having suffered or observed any kind of violence from the teacher in the classroom. These data are especially relevant, since they allow for an overview of violence exposure in countries like Peru, where violence is also structural in many of its regions. Using a large sample of adolescents with different ages, this study confirms the great num- ber of bullying cases observed in recent studies on adolescent population in Peru (see review [22]). Descriptive results also show a greater prevalence of all victimization targets in boys, compared to girls. These data are consistent with most studies that concluded that bullying behavior is more common in boys, who thus suffer more victimization (e.g. [2, 63]). In gender, the most common target is direct-verbal behavior, specifically teasing due to physical appear- ance, and simply to provoke conflict. Other target with high presence is comparison because of academic grades. Both boys and girls also report a high prevalence of other non-typified causes that might be more related to indirect targets, e.g. spreading rumors, cyberbullying, etc., according to Wang's classification [64]. Regarding differences in the prevalence of bully- ing behavior and depression by gender, results are in line with other previous studies, with a higher prevalence of depression in girls and of bullying behavior in boys. According to the first hypothesis, results show that violence relationships in the neighbor- hood enhance violence relationships between peers in school, as well as between teachers and students. These results are extremely relevant, as they indicate the existence of a clear relation- ship between violence perceived in the school environment and the influence of the adoles- cent's main support figures within educational centers. Recent studies already demonstrated the negative relationship between violence in the neighborhood and socio-emotional and respect climate in school [65, 25]. However, some contradictions concerning the influence of Fig 2. Comparison between early and late adolescence SEM models. https://doi.org/10.1371/journal.pone.0174139.g002 PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 9 / 15 Violent relationships: Bullying and depression in early adolescence school environment on bullying behavior still exist in those results [2]. The results of this study show the indirect influence of violence perceived in the school environment on bullying behavior through violence exercised by support figures (peers and teachers). In this sense, results confirm the existence of an indirect effect on the relationship between violent interactions (with peers in school and teachers) and bullying behavior through peer victimization, thereby confirming the second hypothesis. It must be noted that violence from teachers explains better the emergence of aggressive behavior in the adolescent than exposure to the perception of peer violence at school. However, peer victimization is also a predictor of bullying behavior. Former studies characterized bullying victims with a profile different from that of youngsters that are either victims or aggressors (e.g., [66, 67, 68]). Accordingly, results show that being a victim of peers strengthens the relationship between violence from the teacher in the classroom and the adoption of aggressive behaviors. Therefore, if all interactions with significant supports are based on violence, the imitation of these behaviors could inten- sify. Results show that violence exercised by peers and teachers in the school environment explains more than 90% of the total variance explained for bullying behavior. As for the third hypothesis, both violence from peers in school and violence from teachers in the classroom are two clear predictors of depression through loneliness. Nevertheless, this is not true for peer victimization, since it neither shows a predictor effect on loneliness nor on depression after being included into the model. In addition, peer victimization does not con- tribute to a greater explained variance for depression. Consequently, the third hypothesis is only partly confirmed. Results show that violence relationships between the peer groups in school lead to a perception of loneliness in adolescents that, in turn, generates greater depres- sion levels. Therefore, and as pointed out in several studies, social support is a fundamental factor to prevent depression [69, 70]. The same applies to the effect of violence from teachers over depression, in which an indirect effect is also seen through loneliness. In this regard, it must be highlighted the role of the interaction between adolescent students and teachers in the classroom in severe psychological problems such as depression. So, if i nteraction with teach- ers results to be a positive relationship based on confidence, this may generate an important cushion effect when symptoms appear [49] in this stage marked by developmental changes [71]. Finally, differences between the early and late adolescence models confirm the fourth hypothesis. Overall, predictive relationships with victimizations and loneliness and other dependent variables are less strong in late adolescence. In this sense, the literature points out that early adolescence is the most complex stage, since adolescents undergo physical and psy- chological changes and transit from primary to middle school [72, 73]. Therefore, weight should be given to these first stages of transition to adolescence, since they are crucial to pre- vent bullying situations from entrenching. Despite the differences observed between early and late adolescence models, the indirect effects between aggressiveness relationships at home and from teachers, and aggressive behav- ior and depression through mediators remain the same. The model, thus, explains both adoles- cence stages, and the total variance explained by the different variables of the model is almost in the same values for both depression and bullying behavior. This reinforces the need of con- tinuing to adopt effective strategies for preventing bullying, which imply improving affective relationships and strengthening support figures during adolescence. The sample used in this study comprises early and late adolescence, which allowed us to identify differences between these two developmental stages. However, there is an increasing need for longitudinal studies that enable the observation of abiding violence situations in cases of bullying. Consequently, we suggest that future studies address the period from late child- hood to the end of adolescence in order to analyze changes in relationships at a mesosystem PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 10 / 15 Violent relationships: Bullying and depression in early adolescence level, as well as the influence of these changes on the perpetuation of violence. In future research, it would be interesting to assess this phenomenon taking into account the different agents involved in the child's ecological setting and contrast differences by means of multi- level analysis. Conclusions First, the high levels of exposure to violence reported by Peruvian adolescents at both neigh- borhoods where schools are located and within educational centers are noteworthy. An increasing awareness of this issue has been seen in public institutions over the past few years, but interventions are still scarce and insufficient [74]. The prevalence of structural violence in school environments calls for specific measures that allow for preventing violence in a compre- hensive way and from an ecological approach. To this end, a strong coordination among the different services and agents at a community level is necessary for taking ecological preventive actions in which school, family and community work together. Results also show that relation- ships based on violence from significant figures for the adolescent, such as peers and teachers, contribute to the emergence of aggressive behavior and depressive symptoms. In this sense, teachers are the significant adults that should act as positive and stable socio-emotional sup- ports inside schools, and assist in preventing peer conflict during adolescence. The aforemen- tioned requires teachers to be specifically trained to deal with violence situations and to establish effective strategies to improve classroom climate and, in turn, enhance students' aca- demic performance [75, 76]. Otherwise, the prevalence of bullying situations will increase, thus causing the reproduction of violent behavior patterns and the ultimate entrenchment of violence. Acknowledgments We would like to thank the Ministry of Education of Peru (MINEDU) for providing us with the information concerning the baseline of the impact assessment ªSõseVeº, whose data have been analyzed in this work. Author Contributions Conceptualization: XO HCA AA. Formal analysis: XO JTV HCA RM. Funding acquisition: XO. Investigation: RM. Project administration: XO JTV HCA. Supervision: XO. Visualization: XO JTV HCA AA. Writing ± original draft: XO AA MCM JTV. Writing ± review & editing: XO AA. References 1. Bronfenbrenner U. The ecology of human development: Experiments by design and nature. Cam- bridge, MA: Harvard University Press. 1979. PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 11 / 15 Violent relationships: Bullying and depression in early adolescence 2. Hong JS, Espelage DL A review of research on bullying and peer victimization in school: An ecological system analysis. Aggress Violent Behav. 2012; 17(4):311±322. 3. Miller S, Gorman-Smith D, Sullivan T, Orpinas P, Simon TR. Parent and peer predictors of physical dat- ing violence perpetration in early adolescence: Tests of moderation and gender differences. J Clin Child Adolesc Psicol. 2009; 38(4):538±550. 4. Bronfenbrenner U, Morris PA. The bioecological model of human development. In Lerner R.M. & Damon W. R. (Ed.), Handbook of Child Psychology Theoretical Models of Human Development. Hobo- ken, NJ: Wiley; 2006. p. 793±828. 5. Tudge JRH, Mokrova I, Hatfield BE, Karnik RB. Uses and misuses of Bonfenbrenner's bioecological theory of human development. J Fam Theory Rev. 2009; 1(4): 198±210. 6. Hamre BK, Pianta RC. Classroom environments and developmental processes: Conceptualization and measurement. In Meece J. & Eccles J. (Ed.), Handbook of Research on Schools, Schooling, and Human Development. New York, NY: Routledge; 2010. p. 25±41. 7. Hatzenbuehler ML, Duncan D, Johnson R. Neighborhood-level LGBT hate crimes and bullying among sexual minority youths: a geospatial analysis. Violence Vict. 2015; 30(4):663±675. https://doi.org/10. 1891/0886-6708.VV-D-13-00166 PMID: 26160063 8. Espelage DL, Hong JS, Rao MA, Thornberg R. Understanding ecological factors associated with bully- ing across the elementary to middle school transition in the United States. Violence Vict. 2015; 30 (3):470±487. PMID: 26118267 9. Rose CA, Nickerson AB, Stormont M Advancing bullying research from a social-ecological lens: An introduction to the special Issue. School Psych Rev. 2015; 44(4):339±352. 10. Pianta RC, Hamre B, Stuhlman M Relationships between teachers and children. In Reynolds WB, Miller GE, Weiner IB, editors. Handbook of Psychology: Educational Psychology. Hoboken: NJ: Wiley; 2003. pp. 199±234 11. Setanni M, Longobardi C, Sclavo E, Fraire M, Prino LE. Development and psychometric analysis of the student±teacher relationship scale-short form. Front Psychol. 2015; 6:898. https://doi.org/10.3389/ fpsyg.2015.00898 PMID: 26167156 12. Stuhlman MW, Pianta RC Teachers'narratives about their relationships with children: associations with behavior in the classrooms. School Psych Rev. 2002; 31(2):148±163. 13. Parson T. The school class as a social system. Schools and society: A sociological approach to educa- tion, In Ballantine J. Spade J, editors. School and Society. A Sociological Approach to Education US: Sage; 2007. pp. 80±85. 14. Ferguson YL, Kasser T, Jahng S. Differences in life satisfaction and school satisfaction among adoles- cents from three nations: The role of perceived autonomy support. J Res Adolesc. 2011; 21(3):649± 15. Huebner ES, Hills KJ, Jiang X, Long RF, Kelly R, Lyons MD. Schooling and children's subjective well- being. In Ben-Arieh A, Casas F, Frones I, Korbin JE, editors. Handbook of child well-being. Nether- lands: Springer; 2014. pp. 797±819. 16. Schmidt CJ, Pierce J, Stoddard SA. The mediating effect of future expectations on the relationship between neighborhood context and adolescent bullying perpetration. J Community Psychol. 2016; 44 (2):232±248. 17. Jennings WG, Maldonado-Molina MM, Reingle JM, Kromo KA. A multi-level approach to investigating neighborhood effects on physical aggression among urban Chicago youth Am J Crim Justice. 2011; 36 (4):392±407. https://doi.org/10.1007/s12103-011-9118-2 PMID: 24049432 18. Galtung J. Cultural violence. J Peace Res. 1990; 27(3):291±305. 19. World Bank. World development indicators (2013) [Data set]. 2015. Available from http://data. worldbank.org/indicator/SL.TLF.TOTL.IN/countries/USCN?order=wbapi_data_value_2013%20wbapi_ data_value%20wbapi_data_value-last&sort=asc&display=default 20. Borba CPC, Gelaye B, Zayas L, Ulloa M, Lavelle J, Mollica RF, Henderson DC. Making strides towards better mental health care in Peru: results from a primary care mental health training. Int J Clin Psychiatry Ment Health. 2015; 3(1):9±19. https://doi.org/10.12970/2310-8231.2015.03.01.3 PMID: 27054141 21. Thorne C, Corveleyn J, Pezo del Pino C, Vela  zquez T, Valdez R. Buenas pra  cticas en la prevencio  n y atencio  n de la violencia social: Sistematizacio  n de la labor realizada por instituciones y organizaciones peruanas. Lima: PUCP; 2011. 22. Merino C, Carozzo J, Benites L. Bullying in Peru: A code of silence. In: Jimerson SR, Nickerson AB, Mayer MJ, & Furlong M.J, editors. Handbook of School Violence and School Safety International Research and Practice. New York: Routledge; 2011. pp. 153±164. 23. Seinfeld J, Galarza F. Comprendiendo el consumo de bebidas alcoholicas por menores de edad en el Peru  : determinantes de su frecuencia e intensidad. Economõ Âa. 2014; 37(73):47±74. PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 12 / 15 Violent relationships: Bullying and depression in early adolescence 24. Willits D, Broidy L, Denman K. Schools, Neighborhood Risk Factors, and Crime. Crime and Delin- quency. 2013; 59, 292±315. 25. Mann MJ, Kristjansson AL, Sigfusdottir ID, Smith ML. The role of Community, family, peer, and school factors in group bullying: Implications for school-based intervention. J Sch Health. 2015; 85(7):477± 486. https://doi.org/10.1111/josh.12270 PMID: 26032278 26. Sebastian C, Burnett S, Blakemore SJ. Development of the self-concept during adolescence. Trends Cogn. Sci. 2008; 12(11): 441±446. https://doi.org/10.1016/j.tics.2008.07.008 PMID: 18805040 27. Brown B. Adolescent's relationships with peers. In: Lerner R, & Steinberg L, editores. Handbook of Ado- lescent Psychology. Hoboken, NJ: Wiley; 2004. pp. 363±394. 28. Sentse M, Kretschmer T, Salmivalli C. The Longitudinal Interplay between Bullying, Victimization, and Social Status: Age-related and Gender Differences. Soc Dev 2015; 24(3), 659±677. 29. Hay C, Meldrum R. Bullying victimization and adolescent self-harm: Testing hypotheses from general strain theory. J Youth Adolesc. 2010; 39(5),446±459. https://doi.org/10.1007/s10964-009-9502-0 PMID: 20072852 30. Eisenberg M E, Aalsma MC. Bullying and peer victimization: Position paper of the Society for Adoles- cent Medicine. J Youth Adolesc. 2005; 36(1):88±91. 31. Olweus D, Breivik K. Plight of victims of school bullying: The opposite of well-being. In: Ben-Arieh A, Casas F, Frones I, Korbin JE, editors. Handbook of child well-being Netherlands: Springer; 2014. p. 2593±2616. 32. Malecki CK, Demaray MK, Coyle S, Geosling R, Rueger SY, Becker LD. Frequency, power differential, and intentionality and the relationship to anxiety, depression, and self-esteem for victims of bullying. Child Youth Care Forum. 2015; 44(1):115±131. 33. Espelage DL, Bosworth K, Simon TR. Short-term stability and prospective correlates of bullying in mid- dle-school students: An examination of potential demographic, psychosocial, and environmental influ- ences. Violence Vict. 2001; 16(4): 411±426. PMID: 11506450 34. Espelage DL, Swearer SM. Research on school bullying and victimization: What have we learned and where do we go from here? School Psych Rev. 2003; 32(3):365±384. 35. Kaltiala-Heino R, Fro È jd S, Marttunen M. Involvement in bullying and depression in a 2-year follow-up in middle adolescence. Eur Child Adolesc Psychiatry. 2010; 19(1):45±55. https://doi.org/10.1007/s00787- 009-0039-2 PMID: 19588185 36. Este  vez E, Martõ Ânez B, Moreno D, Musitu G. Relaciones familiares, rechazo entre iguales y violencia escolar. Cultura y Educacio  n. 2006; 18(3±4):335±344. 37. Stein JA, Dukes RL, Warren JI. Adolescent male bullies, victims, and bully-victims: A comparison of psychosocial and behavioral characteristics. J Pediatr Psychol. 2007; 32(3): 273±282. https://doi.org/ 10.1093/jpepsy/jsl023 PMID: 16896194 38. Woods S, White E. The association between bullying behaviour, arousal levels and behaviour prob- lems. J Adolesc. 2005; 28(3):381±395. https://doi.org/10.1016/j.adolescence.2004.09.002 PMID: 39. Hartley M, Bauman S, Nixon C, Davis S. Comparative study of bullying victimization among students in general and special education. Except Child. 2015; 81(2):176±193. 40. Shaver P, Furman W, Buhrmester D. Transition to college: Network changes, social skills, and loneli- ness. In Shaver P, Furman W, Buhrmester D, Duck S, Perlman D, editors. Understanding personal rela- tionships: An interdisciplinary approach. Thousand Oaks, CA, US: Sage Publications, Inc.; 1985. pp. 193±219. 41. Wang H, Zhou X, Lu C, Wu J, Deng X, Hong L, He Y. Adolescent bullying involvement and psychosocial aspects of family and school life: A cross-sectional study from Guangdong Province in China. PLoS ONE. 2012; 7(7):e38619. https://doi.org/10.1371/journal.pone.0038619 PMID: 22815693 42. Hughes J, Luo W, Kwok O, Loyd L. Teacher-student support, effortful engagement, and achievement: A 3-year longitudinal study. J Educ Psicol. 2008; 100(1):1±14. 43. OÂConnor E. (2010). Teacher±child relationships as dynamic systems. J Educ Psychol. 2010; 48 (3):187±218. 44. Ladd G, Birch S, Buhs E. Children's Social and Scholastic Lives in Kindergarten: Related Spheres of Influence?. Child Dev. 1999; 70(6):1373±1400. PMID: 10621962 45. Graziano PA, Reavis RD, Keane SP, Calkins SD. The role of emotion regulation in children's early aca- demic success. J Sch Psychol. 2007; 45(1):3±19. https://doi.org/10.1016/j.jsp.2006.09.002 PMID: 46. Allen M, Witt PL, Wheeless LR. The role of teacher immediacy as a motivational factor in student learn- ing: Using meta-analysis to test a causal model. Commun Educ. 2006; 55(1): 21±31. PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 13 / 15 Violent relationships: Bullying and depression in early adolescence 47. Maldonado-Carreño C, Votruba-Drzal E. Teacher-child relationships and the development of academic and behavioral skills during elementary school: A within-and between-child analysis. Child Dev. 2011; 82(2):601±616. https://doi.org/10.1111/j.1467-8624.2010.01533.x PMID: 21410918 48. Wentzel KR, Battle A, Russell SL, Looney LB. Social supports from teachers and peers as predictors of academic and social motivation. Contemp Educ Psychol. 2010; 35(3), 193±202. 49. Wang MT, Brinkworth M, Eccles J. Moderating effects of teacher±student relationship in adolescent tra- jectories of emotional and behavioral adjustment. Dev Psychol. 2013; 49(4):690±705. https://doi.org/ 10.1037/a0027916 PMID: 22582833 50. Seals D, Young J. Bullying and victimization: Prevalence and relationship to gender, grade level, ethnic- ity, self-esteem, and depression. Adolesc. 2003; 38(152): 735±747. 51. Oldehinkel AJ, Veenstra R, Ormel J, De Winter AF, Verhulst FC. Temperament, parenting, and depres- sive symptoms in a population sample of preadolescents. J Child Psychol Psychiatry. 2006; 47(7):684± 695. https://doi.org/10.1111/j.1469-7610.2005.01535.x PMID: 16790003 52. Ormel J, Oldehinkel AJ, Ferdinand RF, Hartman CA, De Winter AF, Veenstra R, Vollebergh W, Mind- eraa RB, Buitelaar JK, Verhulst FC. Internalizing and externalizing problems in adolescence: general and dimension-specific effects of familial loadings and preadolescent temperament traits Psychol Med. 2005; 35(12): 1825±1835. https://doi.org/10.1017/S0033291705005829 PMID: 16300695 53. Ministerio de Educacio  n del Peru  . Ana  lisis psicome  trico del Cuestionario de Bienestar Escolar de Pri- maria y SecundariaÐEstudio Piloto 2013, MINEDU; 2013. 54. Ministerio de Educacio  n del Peru  . Ana  lisis psicome  trico del Cuestionario de Bienestar Escolar de Pri- maria y Secundaria.- Estudio 2014. MINEDU; 2014. 55. Espelage DL, Holt MK. Bullying and victimization during early adolescence: Peer influences and psy- chosocial correlates. J Aggress Maltreat Trauma. 2001; 2(2±3): 123±142. 56. Hays RD, DiMatteo MR. A short-form measure of loneliness. J Pers Assess. 1987; 51(1):69±81. https://doi.org/10.1207/s15327752jpa5101_6 PMID: 3572711 57. Hughes ME, Waite LJ, Hawkley LC, Cacioppo JT. A short scale for measuring loneliness in large sur- veys results from two population-based studies. Res Aging. 2004; 26(6):655±672. https://doi.org/10. 1177/0164027504268574 PMID: 18504506 58. Rebelez J. L., & Furlong M. J. (2013). Psychometric support for an abbreviated version of the California School Climate and Safety Survey. International Journal of School & Educational Psychology, 1(3), 154±165. 59. Bradley KL, Bagnell AL, Brannen CL. Factorial validity of the Center for Epidemiological Studies Depression 10 in adolescents. Issues Ment Health Nurs. 2010; 31(6), 408±412. https://doi.org/10.3109/ 01612840903484105 PMID: 20450343 60. Bojorquez I, Salgado N. Caracterõ Âsticas psicome  tricas de la Escala Center for Epidemiological Studies- depression (CES-D), versiones de 20 y 10 reactivos en mujeres de una zona rural mexicana. Salud Ment. 2009; 32(4): 299±307. 61. Nunnally J. C., & Bernstein I. H. Psychometric theory. New York: McGraw-Hill; 1994. 62. Podsakoff P, MacKenzie S, Lee J, Podsakoff NP. Common method variance in behavioral research: A critical review of the literature and recommended remedies. J of Applied Psychol. 2003; 88 (5): 879± 63. Varjas K, Henrich CC, Meyers J. Urban middle school students' perceptions of bullying, cyberbullying, and school safety. J Sch Violence. 2009; 8(2): 159±176. 64. Wang J, Nansel TR, Iannotti RJ. Cyber and traditional bullying: Differential association with depression. J Adolesc Health. 2011; 48(4):415±417. https://doi.org/10.1016/j.jadohealth.2010.07.012 PMID: 65. McCoy DC, Roy AL, Sirkman GM. Neighborhood crime and school climate as predictors of elementary school academic quality: A cross-lagged panel analysis. American J Community Psychol. 2013; 52(1± 2),128±140. 66. Scholte RH, Overbeek G, Ten Brink G, Rommes E, De Kemp RA, Goossens L, Engels RC. The signifi- cance of reciprocal and unilateral friendships for peer victimization in adolescence. J Youth Adolesc. 2009; 38(1),89±100. https://doi.org/10.1007/s10964-008-9287-6 PMID: 19636794 67. Shin Y. Psychosocial and friendship characteristics of bully/victim subgroups in Korean primary school children. Int J Sch Educ Psicol. 2010; 31(4):372±388. 68. Kochel KP, Ladd GW, Bagwell CL, Yabko BA. Bully/victim Profiles' differential risk for worsening peer acceptance: The role of friendship. J Appl Dev Psicol. 2015; 41,38±45. 69. Hunter SC, Boyle JME. Appraisal and coping strategy use in victims of school bullying. Br J Educ Psicol. 2004; 74(1): 83±107. PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 14 / 15 Violent relationships: Bullying and depression in early adolescence 70. Holt MK, Espelage DL. Perceived social support among bullies, victims, and bully-victims. J Youth Ado- lesc. 2007; 36(8): 984±994. 71. Hanewald R. Transition between primary and secondary school: Why it is important and how it can be supported. Australian Journal of Teacher Education. 2013; 38(1):62±74. 72. Diener E, Oishi S, Lucas RE. Personality, culture, and subjective well-being: Emotional and cognitive evaluations of life. Annu Rev Psychol. 2003; 54(1):403±425. 73. Frey A, Ruchkin V, Martin A, Schwab-Stone M. Adolescents in transition: School and family characteris- tics in the development of violent behaviors entering high school. Child Psychiatry Hum Dev. 2009; 40 (1):1±13. https://doi.org/10.1007/s10578-008-0105-x PMID: 18574687 74. Burela A, Piazza M, Alvarado GF, Gushiken A, Fiestas F. Acceptability of physical punishment in child rearing by people who were victims of physical violence during childhood in Peru. Rev Peru Med Exp Salud Publica. 2014; 31(4):669±675. PMID: 25597716 75. Lo  pez-Gonza  lez L, Amutio A, Oriol X, Bisquerra R. Ha  bitos relacionados con la relajacio  n y la atencio  n plena (mindfulness) en estudiantes de secundaria: influencia en el clima de aula y el rendimiento aca- de  mico. Journal of Psychodidactics. 2016; 21(1):121±138. 76. Oriol X, Lira M, Covarrubias C, Molina V. Positive emotions, Autonomy support, and academic perfor- mance of university students. The mediating role of academic engagement and self-efficacy. Journal of Psychodidactics, 2017; 22(1), 45±53. PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 15 / 15 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png PLoS ONE Pubmed Central

Violent relationships at the social-ecological level: A multi-mediation model to predict adolescent victimization by peers, bullying and depression in early and late adolescence

Loading next page...
 
/lp/pubmed-central/violent-relationships-at-the-social-ecological-level-a-multi-mediation-mNgLVHFfmP

References (81)

Publisher
Pubmed Central
Copyright
© 2017 Oriol et al
ISSN
1932-6203
eISSN
1932-6203
DOI
10.1371/journal.pone.0174139
Publisher site
See Article on Publisher Site

Abstract

Citation: Oriol X, Miranda R, Amutio A, Acosta HC, Mendoza MC, Torres-Vallejos J (2017) Violent relationships at the social-ecological level: A multi- Background mediation model to predict adolescent victimization From the social-ecological perspective, exposure to violence at the different developmental by peers, bullying and depression in early and late adolescence. PLoS ONE 12(3): e0174139. https:// levels is fundamental to explain the dynamics of violence and victimization in educational doi.org/10.1371/journal.pone.0174139 centers. The following study aims at analyzing how these relationships are produced in the Editor: Andrew R. Dalby, University of Peruvian context, where structural violence situations exist. Westminster, UNITED KINGDOM Received: October 30, 2016 Methods Accepted: March 3, 2017 A multi-mediation structural model with 21,416 Peruvian adolescents (M = 13.69; SD = 0.71) was conducted to determine the influence of violence in the school environment on Published: March 30, 2017 violence perceived within school and violence exercised by teachers. In addition, it was also Copyright:© 2017 Oriol et al. This is an open intended to determine whether these violent relationships predict depression through loneli- access article distributed under the terms of the Creative Commons Attribution License, which ness, and bullying through peer victimization. The existence of differences between early permits unrestricted use, distribution, and and late adolescence was also verified. reproduction in any medium, provided the original author and source are credited. Results Data Availability Statement: All relevant data are within the paper. Results confirm that violence in the school setting has high influence on violence exercised by adolescents and teachers within the school. Teacher violence is the most important pre- Funding: This research was made possible with the support of the Ministry of Education of Peru, dictor of depression through loneliness, and encourages peer victimization and the emer- especially the Management Division of Educational gence of aggressive behavior. Exposure to violence exercised by support sourcesÐteachers Quality (MINEDU) who were responsible for and classmatesÐexplains more than 90% of the total variance explained in bullying behav- carrying out the data collection within the ior. Differences were found between early and late adolescence models. framework of the impact assessment carried out by the Specialized System against School Violence (SiseVe). PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 1 / 15 Violent relationships: Bullying and depression in early adolescence Competing interests: The authors have declared Conclusion that no competing interests exist. The high prevalence of structural violence in school settings facilitates the bullying/victimiza- tion dynamics within school. From a social-ecological perspective, this result suggests the importance of network cooperation at a mesosystem level, with teachers from educational centers playing a crucial role in the prevention of bullying/victimization. Introduction From the social-ecological perspective [1], relationships between developmental environments and the dynamics that occur within them are fundamental to understand how the risk of vio- lence, its protective factors, and depression in adolescence are constructed (for revision, [2]). Thus, a central tenet of this theory is that individual development is influenced by the ongoing qualities of the child's social settings and the interactions between these settings (e.g., family, peers, schools, communities) [3]. An educational model oriented towards a conception of the ecological development of the individual cannot help but consider the varied forms of interac- tion and interchange that children experience both inside and outside schools [4, 5, 6]. In the social-ecological model, the mesosystem refers to the relationships between the dif- ferent microsystems. This concept is of paramount importance for studies on bullying [7, 8, 9] and, thus, for this specific study. Specifically, student-teacher relationships are microsystems that consist of the multiple interrelated perceptions that both parties have about their interac- tions [10, 11]. Perceptions are important because they are real, from a psychological stand- point, and they have the power to influence the behavior of each party significantly [12]. During adolescence, the school environment serves a pivotal role in the development pro- cess and asserts itself as one of the most important socialization spaces [13]. In addition, affec- tive relationships are established with peers and adults like teachers, who acquire special significance in this developmental stage [14, 15]. School, as a socializing setting, forms part of a neighborhood or geographical zone which, in turn, influences the experiences and everyday relationships at school [16,11]. Therefore, the study of the structural violence present in zones or neighborhoods where schools are located has become a growing interest over the last years, as it may help explaining bullying phenomena [17, 2, 16]. From the ecological perspective, understanding the influence of diverse factors on victimization/bullying dynamics as well as the relationships between the different actors becomes necessary for adopting effective preven- tive measures. The main factors to be studied are: 1) Contextual factors: According to the Theory of Structural Violence developed by Galtung [18], the extreme adversity conditions to which individuals are subjected within a society may be considered structural violence conditions when they are embedded in the social, political and economic organization of society. Peru is one of the countries where the Shining Path ter- rorist activity from 1980 provoked structural consequences, which makes it be regarded as a country ravaged by violence until today [19]. During decades, many of the population was fre- quently exposed to stress situations that generated important consequences such as depression, anxiety and post-traumatic stress disorders (PTSD) [20]. Currently, violence is considered a pervasive issue in Peru, which manifests in all contexts, such as workplaces, streets, public spaces, and affects equally men, women, adolescents and children [21]. Among Peruvian youngsters, these environments affected by structural violence translate into participation in street gangs and high consumption of alcohol from early ages. Furthermore, 50% of regular consumers start consuming alcohol at age 13, while 90% starts before 16, and this is PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 2 / 15 Violent relationships: Bullying and depression in early adolescence encompassed by a high prevalence of violence in schools [22, 23]. Schools located in settings where structural violence is entrenched might also reproduce certain violence dynamics within the school environment itself as pointed out by recent studies [24, 25]. 2) Relationships with significant support figures: Peer relationships that are mostly devel- oped during the time spent in educational centers come first in the list of adolescents' prior- ities [26, 27]. Adolescents need to be recognized by their peers, and the social support received from the latter is essential to strengthen their self-esteem [28]. Preadolescence and transition to secondary school are especially prone to bullying, due to the physical and psy- chological changes that occur during these periods [29, 2, 30]. Thus, when adolescents are perceived as different from others, whether because of their ethnicity, sex, socioeconomic status, the probabilities of being assaulted by their peers rise [31]. The same is true for psychological variables like depression [32], which may be one of the consequences of vic- timization, since violence, especially from peers, undermines self-esteem and also makes adolescents feel helpless and depressed [33, 34, 35, 32]. Additionally, psychosocial adjust- ment problems and low self-esteem sometimes make victims of abuse become aggressors [36, 37, 38]. Therefore, several studies consider victimization and bullying as part of a same phenomenon [39]. The consequences of bullying may be even more severe if there is a lack of socio-emotional support [40, 2, 41]. In this sense, the affective relationship that teachers establish with students are a key factor for school adaptation [42, 6, 43] and for other variables related to adolescent adjustment, such as social functioning (e.g., [44]), behavior problems (e.g., [45]), and academic achievement [46, 47, 48]. This is confirmed by a recent study conducted by Wang, Brinkworth and Eccles with 1400 youngster, showing that trust and positive affective relationships with teachers act as moderators in the relationship between parent-children conflicts, and depres- sion and bad behavior in 13-to-18-year-old adolescents [49]. In this sense, the type of interac- tions that adolescents establish with significant people, depending on their developmental stage, appears to act as important protective or risk factors for victimization/bullying [3, 9]. Present study The constant changes that adolescents undergone with the onset of puberty makes this a time of special vulnerability, which contributes to the emergence of bullying [2]. From a social-eco- logical standpoint, it is estimated that violence relationships at a mesosystem level may influ- ence the violence interactions that occur between peers during adolescence. In this regard, the structural violence of a context can turn affective relationships into risk factors rather than act- ing as protective factors [28]. Taking all this into consideration, the following hypotheses are proposed: (1) Perception of violence in the neighborhood where a school is located is expected to influence violence between peers within the educational center, as well as violence exercised by teachers in the classroom; (2) Violence relationships (i.e. between peers within the school and from teachers in the classroom) will be predictors of bullying behavior through victimization by peers; (3) Peer violence at school, violence exercised by teachers, and victimization will predict depres- sion through loneliness; and (4) Differences will be found between early and late adolescence in the aforementioned variables. Literature shows differences in aggressiveness and depression according to gender in ado- lescent population. Concretely, most studies report a higher prevalence of bullying cases in boys rather than in girls [50, 2]. Girls, on their part, would have greater levels of negative emo- tionality and depression [51, 52]. Therefore, gender was controlled for in the proposed struc- tural equation model. PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 3 / 15 Violent relationships: Bullying and depression in early adolescence Table 1. Demographic characteristics of participants (n = 21416). Characteristic n % Gender Male 11528 53.8% Female 9888 46.2% Stage of adolescence Early adolescence 14664 69.1% Late adolescence 6559 30.9% https://doi.org/10.1371/journal.pone.0174139.t001 Method Sample Participants were part of an impact assessment study on school violence developed by the Min- istry of Education of Peru, which sought to reduce the cases of violence in schools. Based on this criterion, Ministry of Education of Peru considered schools for this purpose and contacted directly with them'. In total, 21,416 13-to-17-year-old students in 70 schools from 19 different regions at a national level took part in the study (see Table 1). The mean age was 13.69 (SD = 0.71). As for the period of adolescence (early vs. late), most participants were in early adoles- cence. Thirty percent of the sample came from Lima province (i.e., Lima and Lima Callao), which has the greatest population of Peru (see Table 2). Procedure and ethics statement The Ministry of Education of Peru directly contacted with all educational institutions. Parents and principals were informed of the scope and importance of the study, which lasted four months. All questionnaires were administered in the classroom in the presence of at least one member of the research team. Table 2. Sample distributed by provinces. Province n % Amazonas 28 1% Ancash 931 4.3% Arequipa 604 2.8% Ayacucho 248 1.2% Cajamarca 102 05% Cusco 58 0.3% Huanuco 227 1.1% Ica 3.143 14.7% Junõn 2.655 12.4% La Libertad 1.799 8.4% Lambayeque 1.897 8.9% Lima 4.820 22.5% Lima-Callao 1.586 7.4% Pasco 233 1.1% Piura 376 1.8% Puno 804 3.8% San Martõn 570 2.7% Tacna 1.128 5.3% Tumbes 207 1% https://doi.org/10.1371/journal.pone.0174139.t002 PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 4 / 15 Violent relationships: Bullying and depression in early adolescence Ethical approval for this study was granted by the Ministry of Education of Peru (MINEDU). Parents were requested to sign an informed consent for the application of the questionnaire. To protect the privacy of the students, confidentiality of the questionnaires was guaranteed, and students consent was requested prior to their application. International ethical guidelines for studies with human subjects described in the Nuremberg Code and in the Decla- ration of Helsinki were applied. Measures Reason for being bullied. Based on questions about the reasons for being bullied by other classmates. This scale was extracted from the Single School Well-being Questionnaire-CUBE (in Spanish, Cuestionario Unico de Bienestar EscolarÐCUBE) [53, 54]. This questionnaire was elaborated during the impact assessment carried out by the Ministry of Education of Peru to assess the development of socioemotional skills in educational institutions from Lima Met- ropolitan Area. This scale has 10 items that explore the reason for experiencing bullying (e.g. socioeconomic status). Each question was dichotomous, with 0 accounting for (ªNo), and 1 indicating (ªYesº). Kuder-Richardson reliability was .99. Teacher's aggressive behavior. The Teacher Violence Scale was based on questions about victimization extracted also from CUBE [53, 54]. This scale is composed by five items that enquire the student on violence situations directly (e.g. A teacher threatened to hurt you or beat you) or indirectly involving the teacher as the aggressor (e.g. You saw a teacher insulting other student), and whose range of response goes from 0 (ªNeverº) to 2 (ªTwo or more timesº). CronbachÂsα for the studied sample was .90. Bullying behavior. The Bullying Behavior Scale was based on the CUBE scale. This scale was composed by four items aimed at assessing the experience of the student as an aggressor (e.g. I threatened to hurt or beat other student). The response range goes from 0 (ªNeverº) to 2 (ªTwo or more timesº). CronbachÂsα = .89. Victimization. To measure this construct two scales were used: (i) The School Well-being Questionnaire and (ii) the inventory developed by Espelage and Holt [55]. This eight-item scale assesses the frequency of victimization suffered by the student considering the following categories: physical, verbal, exclusion, and cyberbullying (e.g. One or more students insulted you). The response range is from 0 (ªNeverº) to 2 (ªTwo or more times) CronbachÂsα was = .92. Loneliness. Three-item was used to measure this construct, which is based on a brief ver- sion of the 20-item UCLA Loneliness Scale Test (e.g. You feel isolated from others). The responses for these questions range from 1 (ªrarely or neverº) to 4 (ªMost or all the timeº). The advantage of this shorter version is that its application is more practical and less expensive than that of the original scale. It is worth noticing that this version showed appropriate psycho- metric properties in previous studies [56, 57]. CronbachÂsα = .88. Violence in school environment. Three-item self-constructed scale was used to identify vulnerability to violence in the school surroundings. The items are: (1) There are places in the way to or back from school that I don’t like to go through because I’m afraid that somebody hurts me; (2) In my school many students are in gangs and; (3) Crime and violence in my neighborhood are affecting my school, and whose responses range from 0 (ªNoº) to 4 (ªalwaysº). CronbachÂs α = .71. School violence. Four items was used to assess this variable, based on the abbreviated ver- sion of the California School Climate and Safety Survey [58] (e.g. The students of my school get into fights), and with response categories ranging from 0 (ªNoº) to 4 (ªalwaysº). CronbachÂsα was .77 PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 5 / 15 Violent relationships: Bullying and depression in early adolescence Depression. 10-item Depression Symptoms Test developed by Bradley [59] was included. This test consists of a shorter version of the originally 20-item a CES-D [60]. An example item on the scale is: I felt that everything I did was an effort. Response options range from 1 (ªrarely or neverº) to 4 (ªMost or all the timeº). CronbachÂsα was .79. All scales included in this study achieved the cut-off point of .70 proposed by [61] Statistical analysis Descriptive statistics and correlations were calculated by means of the statistical program SPSS 22.0. The statistical package AMOS 18.0 was used for structural equation models, and multiple indirect effects were calculated using maximum likelihood estimation together with bias-cor- rected confidence interval bootstrap test. This procedure provides an average of the estimates obtained from bootstrap samples and their standard error. To verify the model fit, absolute 2 2 and relative goodness-of-fit indices of the model were used, that is:χ indicators and theχ /gl coefficient, Comparative Fit Index (CFI), Incremental Fit Index (IFI), Root Mean Square Error of Approximation (RMSEA) and Standardized Root Mean Square Residual (SRMR). A multi-group analysis was also run to verify differences between the early adolescence and late adolescence models. Results Descriptive analysis Prevalence of exposure to violence. First, percentages were obtained through the responses to the Likert-type scales. The prevalence of violence in school environment, violence into school, as well as the presence of violent gangs were calculated taking as reference the response option corresponding to ªsometimesº in the answers given to the scales measuring violence. In order to calculate the prevalence of violence suffered by teachers and peers, the answers that reported at least one aggression or abuse during the last month were considered. Results show that 9,209 students (43%) felt unsafe when walking around the school sur- roundings. In addition, 12,721 (59.4%) adolescents reported usually seeing violent behaviors in school and 10,494 (49%) informed that youngsters from school belonged to violent gangs. Furthermore, 8,673 (40.5%) students reported having suffered abuse or insults from their peers at school. Finally, 4,237 adolescents (20.3%) stated having suffered or observed some type of violence from the teacher in the classroom. Table 3 summarizes the reasons for bullying behaviors according to the responses of the dichotomous scale. When comparing gender, men presented higher levels of aggressiveness (M = .44 vs. M = .28; p < .001), whereas women present higher levels of depression (M = 2.90 vs. M = 2.00; p < .001). Correlations. Correlations were performed between all the study variables. As shown in Table 4, all correlations were significant and with values above .15. Measurement and structural indirect effects model Firstly, to verify the fit of the variables into the model, HarmanÂs single factor test was con- ducted [62]. Two models were evaluated: (1) One latent factor model, in which all items were included; and (2) multiple latent factors model, in which the variables considered in this study were included. The results showed an adequate fit to the second model proposed. The model fit is as follows: X²/df = 1.66, df = 199; p < .001; CFI = .94; TLI = .93; SRMR = .04; RMSEA = .04. Subsequently, a first structural equation model was conducted to test the proposed hypoth- eses of this study. In this model, violent relationships in the school were considered as predic- tors of bullying behavior through victimization, and victimization by peers was taken as a PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 6 / 15 Violent relationships: Bullying and depression in early adolescence Table 3. Prevalence of reasons for being bullied in male children (n = 11528) and female children (n = 9888). Male children Female children Targets n % N % χ2 a. Racial differences 975 8 717 7 6.17** b. Geographical differences 514 4 310 3 20.44*** c. Sexual condition 603 5 302 3 55.12*** d. Physical appearance 2434 21 1780 18 20.34*** e. Presence of disabilities 293 3 128 1 38.54*** f. Religious beliefs 806 7 538 5 16.14*** g. Socioeconomic status 1192 10 736 7 43.44*** h. School grades (for both high and low scores) 1872 16 1437 15 5.22* i. Without apparent reason 1917 17 1191 12 71.87*** j. Other reasons 2245 19 1734 18 6.32** * p< .05 ** p< .01 *** p< .001. https://doi.org/10.1371/journal.pone.0174139.t003 predictor of depression through loneliness as suggested in hypothesis two and three. Neverthe- less, it was observed that victimization by peers did not show a predictor effect on depression nor contribute to loneliness with more total variance explained. Therefore, a new model with- out that relationship was calculated. The results showed an adequate fit for the final model: X²/ df = 1.66, df = 199; p < .001; CFI = .94; TLI = .93; SRMR = .04; RMSEA = .04 (Fig 1) All direct effects of the model were significant (see Table 5). Regarding indirect effects with bias-correct confidence intervals, an indirect effect is observed in the relationship between school violence and depression through loneliness (CI = [.021, .134]); p = 0.01); and between school violence and bullying behavior through victimization (CI = [.029, .044]); p = 0.05). Fur- thermore, a significant indirect effect between teacher's aggressive behavior and bullying behavior through victimization (CI = [.350, .459]); p = 0.05), and between teacher's aggressive behavior and depression through loneliness (CI = [.013, .044]); p = 0.01) was identified. The variables included in the model explain 15% the variance of depression and 93% of the full model variance explained. Invariance across adolescence stage To calculate the invariance between the early and late adolescence models, a multi-group anal- ysis was conducted. Significant differences were identified between the unconstrained and Table 4. Correlations and CronbachÂsα between all the variables. Variable 1. 2. 3. 4. 5. 6. 7. 1. Violence in the environment Ð 2. School violence .36*** (.71) 3. Teacher's aggressive behavior .26*** .36*** (.90) 4. Loneliness .24*** .20*** .15*** (.88) 5.Victimization .29*** .38*** .81*** .17*** (.92) 6. Depression .21*** .19*** .17*** .31*** .18*** (.79) 7.Bullying behavior .26*** .35*** .91*** .15*** .86*** .17*** (.89) *** p< .001 https://doi.org/10.1371/journal.pone.0174139.t004 PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 7 / 15 Violent relationships: Bullying and depression in early adolescence Fig 1. Multi-mediation structural equation model (SEM). https://doi.org/10.1371/journal.pone.0174139.g001 measurement weight models (p < .001). Specifically, slightly stronger relationships in path estimates were observed in the early adolescence model, although total variance explained in both models remain similar (Table 6; Fig 2). In the two models, indirect effects through victim- ization and loneliness were significant. Discussion First, descriptive data show that almost half of Peruvian youngsters perceives school's environ- ment as unsafe. In particular, 43% feel unsafe when walking down the streets surrounding school, and 49% reports that students from their school belong to violent gangs. This data is consistent with a high prevalence of peer victimization. Specifically, more than 40% of the Table 5. Standardized direct effects of the model. Dependent Variables Independent Variables Direct path estimate School violence <Ð Violence in the school environment .84***[.823, .886] Teacher's aggressive behavior <Ð Violence in the school environment .31***[.291, .327] Loneliness <Ð School violence .41***[.378, .442] Depression <Ð School violence .09*[.065, .125] Victimization <Ð School violence .08* [.063, .090] Bullying behavior <Ð School violence .13**[.104, .150] Loneliness <Ð Teacher's aggressive behavior .07*[.063, .090] Victimization <Ð Teacher's aggressive behavior .87***[.854, .886] Bullying behavior <Ð Teacher's aggressive behavior .63***[.567, .690] Depression <Ð Teacher's aggressive behavior .17***[.128, .195] Depression <Ð Loneliness .28***[.254, .315] Bullying behavior <Ð Victimization .46***[.405, .528] * p< .05 ** p< .01 *** p< .001. Note: 95% intervals are in brackets https://doi.org/10.1371/journal.pone.0174139.t005 PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 8 / 15 Violent relationships: Bullying and depression in early adolescence Table 6. Invariance between models of early and late adolescence. 2 2 2 Models χ df χ /df Δχ Δdf CFI TLI SRMR RMSEA Model 1 14184 1162 - - .96 .96 .02 .02 Model 2 14305 1190 5.20 120.82* 28 .96 .96 .02 .02 Model 3 14353 1204 5.02 171.22* 43 .96 .96 .02 .02 Model 4 15470 1247 5.40 193.66* 50 .96 .96 .02 .02 * p< .001 Model 1: Unconstrained; Model 2: Measurement weights; Model 3: Structural covariances; Measurement residuals https://doi.org/10.1371/journal.pone.0174139.t006 sample informs having suffered from any type of physical or verbal violence, which widely exceeds data found by other studies reporting a prevalence of 20% (see review [2]). Likewise, almost 60% of students have observed any kind of violence in school, and more than 20% states having suffered or observed any kind of violence from the teacher in the classroom. These data are especially relevant, since they allow for an overview of violence exposure in countries like Peru, where violence is also structural in many of its regions. Using a large sample of adolescents with different ages, this study confirms the great num- ber of bullying cases observed in recent studies on adolescent population in Peru (see review [22]). Descriptive results also show a greater prevalence of all victimization targets in boys, compared to girls. These data are consistent with most studies that concluded that bullying behavior is more common in boys, who thus suffer more victimization (e.g. [2, 63]). In gender, the most common target is direct-verbal behavior, specifically teasing due to physical appear- ance, and simply to provoke conflict. Other target with high presence is comparison because of academic grades. Both boys and girls also report a high prevalence of other non-typified causes that might be more related to indirect targets, e.g. spreading rumors, cyberbullying, etc., according to Wang's classification [64]. Regarding differences in the prevalence of bully- ing behavior and depression by gender, results are in line with other previous studies, with a higher prevalence of depression in girls and of bullying behavior in boys. According to the first hypothesis, results show that violence relationships in the neighbor- hood enhance violence relationships between peers in school, as well as between teachers and students. These results are extremely relevant, as they indicate the existence of a clear relation- ship between violence perceived in the school environment and the influence of the adoles- cent's main support figures within educational centers. Recent studies already demonstrated the negative relationship between violence in the neighborhood and socio-emotional and respect climate in school [65, 25]. However, some contradictions concerning the influence of Fig 2. Comparison between early and late adolescence SEM models. https://doi.org/10.1371/journal.pone.0174139.g002 PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 9 / 15 Violent relationships: Bullying and depression in early adolescence school environment on bullying behavior still exist in those results [2]. The results of this study show the indirect influence of violence perceived in the school environment on bullying behavior through violence exercised by support figures (peers and teachers). In this sense, results confirm the existence of an indirect effect on the relationship between violent interactions (with peers in school and teachers) and bullying behavior through peer victimization, thereby confirming the second hypothesis. It must be noted that violence from teachers explains better the emergence of aggressive behavior in the adolescent than exposure to the perception of peer violence at school. However, peer victimization is also a predictor of bullying behavior. Former studies characterized bullying victims with a profile different from that of youngsters that are either victims or aggressors (e.g., [66, 67, 68]). Accordingly, results show that being a victim of peers strengthens the relationship between violence from the teacher in the classroom and the adoption of aggressive behaviors. Therefore, if all interactions with significant supports are based on violence, the imitation of these behaviors could inten- sify. Results show that violence exercised by peers and teachers in the school environment explains more than 90% of the total variance explained for bullying behavior. As for the third hypothesis, both violence from peers in school and violence from teachers in the classroom are two clear predictors of depression through loneliness. Nevertheless, this is not true for peer victimization, since it neither shows a predictor effect on loneliness nor on depression after being included into the model. In addition, peer victimization does not con- tribute to a greater explained variance for depression. Consequently, the third hypothesis is only partly confirmed. Results show that violence relationships between the peer groups in school lead to a perception of loneliness in adolescents that, in turn, generates greater depres- sion levels. Therefore, and as pointed out in several studies, social support is a fundamental factor to prevent depression [69, 70]. The same applies to the effect of violence from teachers over depression, in which an indirect effect is also seen through loneliness. In this regard, it must be highlighted the role of the interaction between adolescent students and teachers in the classroom in severe psychological problems such as depression. So, if i nteraction with teach- ers results to be a positive relationship based on confidence, this may generate an important cushion effect when symptoms appear [49] in this stage marked by developmental changes [71]. Finally, differences between the early and late adolescence models confirm the fourth hypothesis. Overall, predictive relationships with victimizations and loneliness and other dependent variables are less strong in late adolescence. In this sense, the literature points out that early adolescence is the most complex stage, since adolescents undergo physical and psy- chological changes and transit from primary to middle school [72, 73]. Therefore, weight should be given to these first stages of transition to adolescence, since they are crucial to pre- vent bullying situations from entrenching. Despite the differences observed between early and late adolescence models, the indirect effects between aggressiveness relationships at home and from teachers, and aggressive behav- ior and depression through mediators remain the same. The model, thus, explains both adoles- cence stages, and the total variance explained by the different variables of the model is almost in the same values for both depression and bullying behavior. This reinforces the need of con- tinuing to adopt effective strategies for preventing bullying, which imply improving affective relationships and strengthening support figures during adolescence. The sample used in this study comprises early and late adolescence, which allowed us to identify differences between these two developmental stages. However, there is an increasing need for longitudinal studies that enable the observation of abiding violence situations in cases of bullying. Consequently, we suggest that future studies address the period from late child- hood to the end of adolescence in order to analyze changes in relationships at a mesosystem PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 10 / 15 Violent relationships: Bullying and depression in early adolescence level, as well as the influence of these changes on the perpetuation of violence. In future research, it would be interesting to assess this phenomenon taking into account the different agents involved in the child's ecological setting and contrast differences by means of multi- level analysis. Conclusions First, the high levels of exposure to violence reported by Peruvian adolescents at both neigh- borhoods where schools are located and within educational centers are noteworthy. An increasing awareness of this issue has been seen in public institutions over the past few years, but interventions are still scarce and insufficient [74]. The prevalence of structural violence in school environments calls for specific measures that allow for preventing violence in a compre- hensive way and from an ecological approach. To this end, a strong coordination among the different services and agents at a community level is necessary for taking ecological preventive actions in which school, family and community work together. Results also show that relation- ships based on violence from significant figures for the adolescent, such as peers and teachers, contribute to the emergence of aggressive behavior and depressive symptoms. In this sense, teachers are the significant adults that should act as positive and stable socio-emotional sup- ports inside schools, and assist in preventing peer conflict during adolescence. The aforemen- tioned requires teachers to be specifically trained to deal with violence situations and to establish effective strategies to improve classroom climate and, in turn, enhance students' aca- demic performance [75, 76]. Otherwise, the prevalence of bullying situations will increase, thus causing the reproduction of violent behavior patterns and the ultimate entrenchment of violence. Acknowledgments We would like to thank the Ministry of Education of Peru (MINEDU) for providing us with the information concerning the baseline of the impact assessment ªSõseVeº, whose data have been analyzed in this work. Author Contributions Conceptualization: XO HCA AA. Formal analysis: XO JTV HCA RM. Funding acquisition: XO. Investigation: RM. Project administration: XO JTV HCA. Supervision: XO. Visualization: XO JTV HCA AA. Writing ± original draft: XO AA MCM JTV. Writing ± review & editing: XO AA. References 1. Bronfenbrenner U. The ecology of human development: Experiments by design and nature. Cam- bridge, MA: Harvard University Press. 1979. PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 11 / 15 Violent relationships: Bullying and depression in early adolescence 2. Hong JS, Espelage DL A review of research on bullying and peer victimization in school: An ecological system analysis. Aggress Violent Behav. 2012; 17(4):311±322. 3. Miller S, Gorman-Smith D, Sullivan T, Orpinas P, Simon TR. Parent and peer predictors of physical dat- ing violence perpetration in early adolescence: Tests of moderation and gender differences. J Clin Child Adolesc Psicol. 2009; 38(4):538±550. 4. Bronfenbrenner U, Morris PA. The bioecological model of human development. In Lerner R.M. & Damon W. R. (Ed.), Handbook of Child Psychology Theoretical Models of Human Development. Hobo- ken, NJ: Wiley; 2006. p. 793±828. 5. Tudge JRH, Mokrova I, Hatfield BE, Karnik RB. Uses and misuses of Bonfenbrenner's bioecological theory of human development. J Fam Theory Rev. 2009; 1(4): 198±210. 6. Hamre BK, Pianta RC. Classroom environments and developmental processes: Conceptualization and measurement. In Meece J. & Eccles J. (Ed.), Handbook of Research on Schools, Schooling, and Human Development. New York, NY: Routledge; 2010. p. 25±41. 7. Hatzenbuehler ML, Duncan D, Johnson R. Neighborhood-level LGBT hate crimes and bullying among sexual minority youths: a geospatial analysis. Violence Vict. 2015; 30(4):663±675. https://doi.org/10. 1891/0886-6708.VV-D-13-00166 PMID: 26160063 8. Espelage DL, Hong JS, Rao MA, Thornberg R. Understanding ecological factors associated with bully- ing across the elementary to middle school transition in the United States. Violence Vict. 2015; 30 (3):470±487. PMID: 26118267 9. Rose CA, Nickerson AB, Stormont M Advancing bullying research from a social-ecological lens: An introduction to the special Issue. School Psych Rev. 2015; 44(4):339±352. 10. Pianta RC, Hamre B, Stuhlman M Relationships between teachers and children. In Reynolds WB, Miller GE, Weiner IB, editors. Handbook of Psychology: Educational Psychology. Hoboken: NJ: Wiley; 2003. pp. 199±234 11. Setanni M, Longobardi C, Sclavo E, Fraire M, Prino LE. Development and psychometric analysis of the student±teacher relationship scale-short form. Front Psychol. 2015; 6:898. https://doi.org/10.3389/ fpsyg.2015.00898 PMID: 26167156 12. Stuhlman MW, Pianta RC Teachers'narratives about their relationships with children: associations with behavior in the classrooms. School Psych Rev. 2002; 31(2):148±163. 13. Parson T. The school class as a social system. Schools and society: A sociological approach to educa- tion, In Ballantine J. Spade J, editors. School and Society. A Sociological Approach to Education US: Sage; 2007. pp. 80±85. 14. Ferguson YL, Kasser T, Jahng S. Differences in life satisfaction and school satisfaction among adoles- cents from three nations: The role of perceived autonomy support. J Res Adolesc. 2011; 21(3):649± 15. Huebner ES, Hills KJ, Jiang X, Long RF, Kelly R, Lyons MD. Schooling and children's subjective well- being. In Ben-Arieh A, Casas F, Frones I, Korbin JE, editors. Handbook of child well-being. Nether- lands: Springer; 2014. pp. 797±819. 16. Schmidt CJ, Pierce J, Stoddard SA. The mediating effect of future expectations on the relationship between neighborhood context and adolescent bullying perpetration. J Community Psychol. 2016; 44 (2):232±248. 17. Jennings WG, Maldonado-Molina MM, Reingle JM, Kromo KA. A multi-level approach to investigating neighborhood effects on physical aggression among urban Chicago youth Am J Crim Justice. 2011; 36 (4):392±407. https://doi.org/10.1007/s12103-011-9118-2 PMID: 24049432 18. Galtung J. Cultural violence. J Peace Res. 1990; 27(3):291±305. 19. World Bank. World development indicators (2013) [Data set]. 2015. Available from http://data. worldbank.org/indicator/SL.TLF.TOTL.IN/countries/USCN?order=wbapi_data_value_2013%20wbapi_ data_value%20wbapi_data_value-last&sort=asc&display=default 20. Borba CPC, Gelaye B, Zayas L, Ulloa M, Lavelle J, Mollica RF, Henderson DC. Making strides towards better mental health care in Peru: results from a primary care mental health training. Int J Clin Psychiatry Ment Health. 2015; 3(1):9±19. https://doi.org/10.12970/2310-8231.2015.03.01.3 PMID: 27054141 21. Thorne C, Corveleyn J, Pezo del Pino C, Vela  zquez T, Valdez R. Buenas pra  cticas en la prevencio  n y atencio  n de la violencia social: Sistematizacio  n de la labor realizada por instituciones y organizaciones peruanas. Lima: PUCP; 2011. 22. Merino C, Carozzo J, Benites L. Bullying in Peru: A code of silence. In: Jimerson SR, Nickerson AB, Mayer MJ, & Furlong M.J, editors. Handbook of School Violence and School Safety International Research and Practice. New York: Routledge; 2011. pp. 153±164. 23. Seinfeld J, Galarza F. Comprendiendo el consumo de bebidas alcoholicas por menores de edad en el Peru  : determinantes de su frecuencia e intensidad. Economõ Âa. 2014; 37(73):47±74. PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 12 / 15 Violent relationships: Bullying and depression in early adolescence 24. Willits D, Broidy L, Denman K. Schools, Neighborhood Risk Factors, and Crime. Crime and Delin- quency. 2013; 59, 292±315. 25. Mann MJ, Kristjansson AL, Sigfusdottir ID, Smith ML. The role of Community, family, peer, and school factors in group bullying: Implications for school-based intervention. J Sch Health. 2015; 85(7):477± 486. https://doi.org/10.1111/josh.12270 PMID: 26032278 26. Sebastian C, Burnett S, Blakemore SJ. Development of the self-concept during adolescence. Trends Cogn. Sci. 2008; 12(11): 441±446. https://doi.org/10.1016/j.tics.2008.07.008 PMID: 18805040 27. Brown B. Adolescent's relationships with peers. In: Lerner R, & Steinberg L, editores. Handbook of Ado- lescent Psychology. Hoboken, NJ: Wiley; 2004. pp. 363±394. 28. Sentse M, Kretschmer T, Salmivalli C. The Longitudinal Interplay between Bullying, Victimization, and Social Status: Age-related and Gender Differences. Soc Dev 2015; 24(3), 659±677. 29. Hay C, Meldrum R. Bullying victimization and adolescent self-harm: Testing hypotheses from general strain theory. J Youth Adolesc. 2010; 39(5),446±459. https://doi.org/10.1007/s10964-009-9502-0 PMID: 20072852 30. Eisenberg M E, Aalsma MC. Bullying and peer victimization: Position paper of the Society for Adoles- cent Medicine. J Youth Adolesc. 2005; 36(1):88±91. 31. Olweus D, Breivik K. Plight of victims of school bullying: The opposite of well-being. In: Ben-Arieh A, Casas F, Frones I, Korbin JE, editors. Handbook of child well-being Netherlands: Springer; 2014. p. 2593±2616. 32. Malecki CK, Demaray MK, Coyle S, Geosling R, Rueger SY, Becker LD. Frequency, power differential, and intentionality and the relationship to anxiety, depression, and self-esteem for victims of bullying. Child Youth Care Forum. 2015; 44(1):115±131. 33. Espelage DL, Bosworth K, Simon TR. Short-term stability and prospective correlates of bullying in mid- dle-school students: An examination of potential demographic, psychosocial, and environmental influ- ences. Violence Vict. 2001; 16(4): 411±426. PMID: 11506450 34. Espelage DL, Swearer SM. Research on school bullying and victimization: What have we learned and where do we go from here? School Psych Rev. 2003; 32(3):365±384. 35. Kaltiala-Heino R, Fro È jd S, Marttunen M. Involvement in bullying and depression in a 2-year follow-up in middle adolescence. Eur Child Adolesc Psychiatry. 2010; 19(1):45±55. https://doi.org/10.1007/s00787- 009-0039-2 PMID: 19588185 36. Este  vez E, Martõ Ânez B, Moreno D, Musitu G. Relaciones familiares, rechazo entre iguales y violencia escolar. Cultura y Educacio  n. 2006; 18(3±4):335±344. 37. Stein JA, Dukes RL, Warren JI. Adolescent male bullies, victims, and bully-victims: A comparison of psychosocial and behavioral characteristics. J Pediatr Psychol. 2007; 32(3): 273±282. https://doi.org/ 10.1093/jpepsy/jsl023 PMID: 16896194 38. Woods S, White E. The association between bullying behaviour, arousal levels and behaviour prob- lems. J Adolesc. 2005; 28(3):381±395. https://doi.org/10.1016/j.adolescence.2004.09.002 PMID: 39. Hartley M, Bauman S, Nixon C, Davis S. Comparative study of bullying victimization among students in general and special education. Except Child. 2015; 81(2):176±193. 40. Shaver P, Furman W, Buhrmester D. Transition to college: Network changes, social skills, and loneli- ness. In Shaver P, Furman W, Buhrmester D, Duck S, Perlman D, editors. Understanding personal rela- tionships: An interdisciplinary approach. Thousand Oaks, CA, US: Sage Publications, Inc.; 1985. pp. 193±219. 41. Wang H, Zhou X, Lu C, Wu J, Deng X, Hong L, He Y. Adolescent bullying involvement and psychosocial aspects of family and school life: A cross-sectional study from Guangdong Province in China. PLoS ONE. 2012; 7(7):e38619. https://doi.org/10.1371/journal.pone.0038619 PMID: 22815693 42. Hughes J, Luo W, Kwok O, Loyd L. Teacher-student support, effortful engagement, and achievement: A 3-year longitudinal study. J Educ Psicol. 2008; 100(1):1±14. 43. OÂConnor E. (2010). Teacher±child relationships as dynamic systems. J Educ Psychol. 2010; 48 (3):187±218. 44. Ladd G, Birch S, Buhs E. Children's Social and Scholastic Lives in Kindergarten: Related Spheres of Influence?. Child Dev. 1999; 70(6):1373±1400. PMID: 10621962 45. Graziano PA, Reavis RD, Keane SP, Calkins SD. The role of emotion regulation in children's early aca- demic success. J Sch Psychol. 2007; 45(1):3±19. https://doi.org/10.1016/j.jsp.2006.09.002 PMID: 46. Allen M, Witt PL, Wheeless LR. The role of teacher immediacy as a motivational factor in student learn- ing: Using meta-analysis to test a causal model. Commun Educ. 2006; 55(1): 21±31. PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 13 / 15 Violent relationships: Bullying and depression in early adolescence 47. Maldonado-Carreño C, Votruba-Drzal E. Teacher-child relationships and the development of academic and behavioral skills during elementary school: A within-and between-child analysis. Child Dev. 2011; 82(2):601±616. https://doi.org/10.1111/j.1467-8624.2010.01533.x PMID: 21410918 48. Wentzel KR, Battle A, Russell SL, Looney LB. Social supports from teachers and peers as predictors of academic and social motivation. Contemp Educ Psychol. 2010; 35(3), 193±202. 49. Wang MT, Brinkworth M, Eccles J. Moderating effects of teacher±student relationship in adolescent tra- jectories of emotional and behavioral adjustment. Dev Psychol. 2013; 49(4):690±705. https://doi.org/ 10.1037/a0027916 PMID: 22582833 50. Seals D, Young J. Bullying and victimization: Prevalence and relationship to gender, grade level, ethnic- ity, self-esteem, and depression. Adolesc. 2003; 38(152): 735±747. 51. Oldehinkel AJ, Veenstra R, Ormel J, De Winter AF, Verhulst FC. Temperament, parenting, and depres- sive symptoms in a population sample of preadolescents. J Child Psychol Psychiatry. 2006; 47(7):684± 695. https://doi.org/10.1111/j.1469-7610.2005.01535.x PMID: 16790003 52. Ormel J, Oldehinkel AJ, Ferdinand RF, Hartman CA, De Winter AF, Veenstra R, Vollebergh W, Mind- eraa RB, Buitelaar JK, Verhulst FC. Internalizing and externalizing problems in adolescence: general and dimension-specific effects of familial loadings and preadolescent temperament traits Psychol Med. 2005; 35(12): 1825±1835. https://doi.org/10.1017/S0033291705005829 PMID: 16300695 53. Ministerio de Educacio  n del Peru  . Ana  lisis psicome  trico del Cuestionario de Bienestar Escolar de Pri- maria y SecundariaÐEstudio Piloto 2013, MINEDU; 2013. 54. Ministerio de Educacio  n del Peru  . Ana  lisis psicome  trico del Cuestionario de Bienestar Escolar de Pri- maria y Secundaria.- Estudio 2014. MINEDU; 2014. 55. Espelage DL, Holt MK. Bullying and victimization during early adolescence: Peer influences and psy- chosocial correlates. J Aggress Maltreat Trauma. 2001; 2(2±3): 123±142. 56. Hays RD, DiMatteo MR. A short-form measure of loneliness. J Pers Assess. 1987; 51(1):69±81. https://doi.org/10.1207/s15327752jpa5101_6 PMID: 3572711 57. Hughes ME, Waite LJ, Hawkley LC, Cacioppo JT. A short scale for measuring loneliness in large sur- veys results from two population-based studies. Res Aging. 2004; 26(6):655±672. https://doi.org/10. 1177/0164027504268574 PMID: 18504506 58. Rebelez J. L., & Furlong M. J. (2013). Psychometric support for an abbreviated version of the California School Climate and Safety Survey. International Journal of School & Educational Psychology, 1(3), 154±165. 59. Bradley KL, Bagnell AL, Brannen CL. Factorial validity of the Center for Epidemiological Studies Depression 10 in adolescents. Issues Ment Health Nurs. 2010; 31(6), 408±412. https://doi.org/10.3109/ 01612840903484105 PMID: 20450343 60. Bojorquez I, Salgado N. Caracterõ Âsticas psicome  tricas de la Escala Center for Epidemiological Studies- depression (CES-D), versiones de 20 y 10 reactivos en mujeres de una zona rural mexicana. Salud Ment. 2009; 32(4): 299±307. 61. Nunnally J. C., & Bernstein I. H. Psychometric theory. New York: McGraw-Hill; 1994. 62. Podsakoff P, MacKenzie S, Lee J, Podsakoff NP. Common method variance in behavioral research: A critical review of the literature and recommended remedies. J of Applied Psychol. 2003; 88 (5): 879± 63. Varjas K, Henrich CC, Meyers J. Urban middle school students' perceptions of bullying, cyberbullying, and school safety. J Sch Violence. 2009; 8(2): 159±176. 64. Wang J, Nansel TR, Iannotti RJ. Cyber and traditional bullying: Differential association with depression. J Adolesc Health. 2011; 48(4):415±417. https://doi.org/10.1016/j.jadohealth.2010.07.012 PMID: 65. McCoy DC, Roy AL, Sirkman GM. Neighborhood crime and school climate as predictors of elementary school academic quality: A cross-lagged panel analysis. American J Community Psychol. 2013; 52(1± 2),128±140. 66. Scholte RH, Overbeek G, Ten Brink G, Rommes E, De Kemp RA, Goossens L, Engels RC. The signifi- cance of reciprocal and unilateral friendships for peer victimization in adolescence. J Youth Adolesc. 2009; 38(1),89±100. https://doi.org/10.1007/s10964-008-9287-6 PMID: 19636794 67. Shin Y. Psychosocial and friendship characteristics of bully/victim subgroups in Korean primary school children. Int J Sch Educ Psicol. 2010; 31(4):372±388. 68. Kochel KP, Ladd GW, Bagwell CL, Yabko BA. Bully/victim Profiles' differential risk for worsening peer acceptance: The role of friendship. J Appl Dev Psicol. 2015; 41,38±45. 69. Hunter SC, Boyle JME. Appraisal and coping strategy use in victims of school bullying. Br J Educ Psicol. 2004; 74(1): 83±107. PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 14 / 15 Violent relationships: Bullying and depression in early adolescence 70. Holt MK, Espelage DL. Perceived social support among bullies, victims, and bully-victims. J Youth Ado- lesc. 2007; 36(8): 984±994. 71. Hanewald R. Transition between primary and secondary school: Why it is important and how it can be supported. Australian Journal of Teacher Education. 2013; 38(1):62±74. 72. Diener E, Oishi S, Lucas RE. Personality, culture, and subjective well-being: Emotional and cognitive evaluations of life. Annu Rev Psychol. 2003; 54(1):403±425. 73. Frey A, Ruchkin V, Martin A, Schwab-Stone M. Adolescents in transition: School and family characteris- tics in the development of violent behaviors entering high school. Child Psychiatry Hum Dev. 2009; 40 (1):1±13. https://doi.org/10.1007/s10578-008-0105-x PMID: 18574687 74. Burela A, Piazza M, Alvarado GF, Gushiken A, Fiestas F. Acceptability of physical punishment in child rearing by people who were victims of physical violence during childhood in Peru. Rev Peru Med Exp Salud Publica. 2014; 31(4):669±675. PMID: 25597716 75. Lo  pez-Gonza  lez L, Amutio A, Oriol X, Bisquerra R. Ha  bitos relacionados con la relajacio  n y la atencio  n plena (mindfulness) en estudiantes de secundaria: influencia en el clima de aula y el rendimiento aca- de  mico. Journal of Psychodidactics. 2016; 21(1):121±138. 76. Oriol X, Lira M, Covarrubias C, Molina V. Positive emotions, Autonomy support, and academic perfor- mance of university students. The mediating role of academic engagement and self-efficacy. Journal of Psychodidactics, 2017; 22(1), 45±53. PLOS ONE | https://doi.org/10.1371/journal.pone.0174139 March 30, 2017 15 / 15

Journal

PLoS ONEPubmed Central

Published: Mar 30, 2017

There are no references for this article.