The Complex Nature of School Violence: Attitudes Toward Aggression, Empathy and Involvement Profiles in Violence

Introduction

In recent years, aggressive adolescent behavior has become a subject of interest in education, as the transition and adaptation to secondary education (in Spain, grade levels in this stage include 12-to-16-year-olds, and up to 18 if a year is repeated)¹ can generate significant student behavior problems² related to social,³ family and individual factors.^4–6 Aggressive behavior involves violent conduct intended to physically or verbally harm or hurt others.^7,8 Aggression may be an impulsive defensive instinct, motivated by anger and frustration, from a threat or conflict,^9–12 or premeditated as part of planned attacks for the specific purpose of obtaining a reward as the result.¹³ There are different types of involvement in school violence within the framework of aggression, usually verbal in adolescents,¹⁴ in which one may have the role of victim, aggressor or observer. Adolescent victims of violence by other students usually have problems with interiorization, are introverted, and do not belong to a peer group. Youths who are the object of aggression may become emotionally distressed by the physical and psychological violence they are subjected to, as well as social exclusion and rejection by other students, increasing the probability of depressive symptomatology, suicidal ideation, stress, anxiety, substance use,^15–18 and psychological adjustment problems.¹⁹ A victim-aggressor profile, in which the adolescent experiences episodes of violence, and in turn, attacks other students, instigating confrontation, has also been found.²⁰

Another role in school violence is the aggressor, whose behavior may be associated with low levels of social and emotional competence,²¹ as well as lower empathy than shown by the victims.²² Often adolescents who inflict violence on others do so because they themselves are suffering or have suffered from abuse or early abandonment, or whose behavior is derived from living in a conflictive setting within the family, or where peers negatively influence their behavior.^23,24

Youths who develop in a violent setting show more stable aggressive traits and belief in the use of aggression. This, along with patterns of upbringing marked by rejection, severity or overprotection can cause moral disconnection from their acts and lead them to stronger justification of the use of violence, or to attribute the reason for such behavior to factors outside themselves.²⁵ As a result, many adolescent aggressors do not recognize having used violence on other students, and only a small percentage admit to such conduct, while a larger percentage of victims of school violence do recognize having been the subject of violence by another student.^26,27 Moreover, according to the meta-analysis by Zych et al²⁸ bullying peers is related to violent behavior in couple relations, and could be different manifestations of the same antisocial dispositions.

Observers, who view student aggression from different perspectives, may also be involved in violence, by either encouraging the aggressors to continue with their violence against other students or help and support the victims.²⁹ In another position, observers do not want to become involved because they fear repercussions to themselves if they intervene to resolve the conflict.^30–32 In this regard, a study by Bauman et al³³ showed that youths who witness victimization of other students and do decide to intervene are against bullying and have more empathy. Cuevas and Marmolejo³⁴ found the role of observer to be of special importance, because it contributes to violent episodes with intimidating behavior, or on the contrary, actions for stopping them.

Factors Involved in Violent Attitudes

Attitude toward violence and deviant behavior, has been shown to be a significant variable in predicting violent conduct of youths.³⁵ One of the factors linked to the formation of attitudes toward violence in adolescents is the influence of the peer group.^36,37 The presence of conflictive, violent students increases the possibility of others supporting and developing disruptive behavior.³⁸ Youths who condone violent conduct often do so to gain acceptance and support from a group of conflictive peers.^39,40

Nevertheless, high-quality friendships favor less violent behavior, and function as a socially and emotionally beneficial protective factor.⁴¹ Attachment relationships with a peer group that provides them with support and security³⁷ are fundamental to diminishing violent behaviors in aggressive young people.⁴²

Parents are another determining factor in developing attitudes favoring or opposing use of violence. The family usually influences adolescent behavior and beliefs including levels of acceptance of violence in both boys and girls and, thereby determining to a greater or lesser extent the likelihood of their developing violent behavior.^43–46 Thus, parents who positively reinforce their children by giving them support and advice on how to handle the conflicts they must confront, promote alternatives to violence for solving their problems.^47,48 In an environment of positive family relations, determined by attachment, inductive, open, positive and empathic communication, the possibility of adolescents having aggressive conduct diminishes.^49–52

Sometimes aggressive adolescent behavior is associated with a conflictive parental relationship and inadequate parenting.^53–55 Lack of family communication, absence of parental authority and low perception of parental support are associated with high levels of adolescent violence.^45,56–58

Furthermore, parental approval of violence as a way of attaining goals proposed and frequent exposure to violence in the family setting cause adolescents to ignore the negative consequences of violence, normalizing aggressive action.^59,60

The role of physical activity in violent behavior during adolescence should also be mentioned, having found a direct positive relationship between affective empathy and cognitive empathy, with differences between sedentary adolescents and those who usually practice some physical activity.⁶¹

Relationship Between Empathy and School Violence

In view of the repercussions that adolescent violent behavior could have, the individual factors that could lead young people to become a victim, aggressor or observer must be known.⁶²

Among the individual factors, empathy has an essential role in violent conduct. This variable can be defined in terms of its two dimensions as cognitive empathy, or the ability to understand the mental states and feelings of others, and affective empathy, or being able to feel them.⁶³ According to the study by Euler et al⁶⁴ and Hartmann et al⁶⁵ high empathy is related negatively to aggressive behaviors, and reduces unjustified violent behavior.^66–68 In line with this, high levels of empathy also act as a protective factor against appearance of violent behavior and bullying, while low levels would be a risk agent.^69–71 A high level of empathy also predicts parental and peer group support, which means an increase in the ability to resolve conflicts, reducing violent behavior.⁷² The study by Eisenberg et al⁷³ further shows high levels of aggressiveness in youths to be related to low empathy derived from the poor relationship and conflict with parents and other students.⁷⁴

A review of the literature on the subject demonstrated the increase in studies on violence in the school context, as well as the influence of empathy on violent conduct. Thus, this study sought to find out more about aggression and its link with empathy and the attitude of youths and their parents toward the use of violence.

The specific objectives of this study were to:

- Analyze the relationships between adolescent and parental attitudes toward aggression and empathy.

- Identify profiles of direct and indirect involvement in school violence.

- Determine whether there are between-group differences in the components of empathy and attitudes toward aggression.

Based on these objectives, the following hypotheses were posed: First, there is a positive association between favorable beliefs about the use of aggression and parental support for violence as a way to resolve conflicts. Thus, youths with more favorable beliefs about aggression would perceive that their parents support violence to a greater extent as a strategy in conflicts, while those with beliefs favoring the use of nonviolent strategies would be more empathic and perceive more parental support toward the use of pacific strategies (H1). We also expected to find different profiles with regard to direct or indirect involvement in school violence, where those who do not use violence toward other students and those who act when they are witness to this type of behavior would be those who would show the most empathy (H2).

Materials and Methods

Participants

A total of 1287 students at public high schools in the province of Almeria (Spain), aged 14 to 18 and with a mean age of 15.11 (SD=0.91), of whom 55% (n=707) were in 11th grade and 45% (577) in 12th grade, participated in this study. The sex distribution was 47.1% (n= 606) boys and 52.9% (n= 681) girls, with mean ages of 15.2 (SD=0.94) and 15.10 (SD= 0.88), respectively.

Instruments

Beliefs about Aggression and Alternatives:⁷⁵ The questionnaire is comprised of a total of 12 items measuring two dimensions: adolescent beliefs about the use of aggression and approval of nonviolent response to hypothetical situations. There are seven items evaluating beliefs about aggression and another five on the use of nonviolent strategies rated on a four-point scale: 1 = Strongly agree, 2 = Agree somewhat, 3 = Disagree somewhat, 4 = Strongly disagree. The reliability found for beliefs about aggression was ω=0.787 and the GLB=0.795, and for nonviolent strategies it was ω=0.555, and GLB=0.602.

Parental Support for Fighting Scale:⁷⁶ This is made up of 10 items which evaluate parental attitudes toward violence on two scales: strong support from parents for use of aggression in response to conflict, and strong support for use of nonviolent strategies by adolescents. There are five items with yes or no answer choices indicating strong parental support for aggression in response to conflict and five showing strong parental support for pacific response. Reliability found for the aggressive solution scale was ω= 0.633 and GLB= 0.689, and for non-aggressive solutions ω=0.653 and GLB= 0.676.

Interpersonal Reactivity Index.^77,78 The Spanish version by Mestre et al⁷⁹ was employed. This scale is comprised of a total of 28 items which measure four dimensions of global empathy: Perspective Taking, Fantasy, Empathic Concern and Personal Distress. This study used two of the four dimensions (Perspective Taking and Empathic Concern), as representative of cognitive and affective empathy. Perspective Taking (PT) refers to complex sociocognitive development enabling one to identify and understand another’s point of view. Within cognitive empathy, Perspective Taking further involves understanding the thoughts or beliefs of others.⁸⁰ Reliability in this study was ω= 0.553 and GLB=0.648. Empathic Concern (EC), the emotional component of empathy, which is the ability of an individual to perceive and identify with feelings of concern and personal distress of those who experience complex harmful situations, had a reliability index of ω=0.542 and GLB=0.615. The items are rated on a five-point Likert-type scale (from 0 to 4), where 0=Does not describe me well, 1=Describes me a little, 2=Describes me well, 3=Describes me rather well, and 4=describes me very well.

Procedure

Before starting to collect data, the school principals were contacted and meetings were set up to inform them of the study objectives. They were also guaranteed confidentiality in data processing. Following the schedule of data collection sessions, two members of the research team went to the schools to implement the questionnaires. The validated tests were administered during the first half of 2019 in the usual classroom assigned to each group, in all cases, in the presence of their teacher/counselor. At the beginning of the session, the students were given the instructions for filling in the questionnaires, with time to solve any questions they might have about it, and they were assured that their answers would be anonymous, and therefore, their privacy would be respected in statistical data processing. The students filled out the tests individually in an estimated mean time of 25–30 minutes. In all cases, the ethical standards of research in the Declaration of Helsinki were complied with. Informed consent was obtained from parents/guardians and also from the participants themselves. The study was approved by the Bioethics Committee of the University of Almeria (Ref: UALBIO2018/015).

Data Analysis

First the bivariate correlation matrix and descriptive statistics were calculated to establish relationships between variables. For interpretation of the magnitude of correlation coefficients, following Cohen,⁸¹ the absolute value of the coefficient was taken, regardless of the sign, such that: r < 0.10 null correlation, 0.10 ≤ r < 0.30 low, 0.30 ≤ r < 0.50 medium, 0.50 ≤ r < 1.00 large.

A two-stage cluster analysis was performed to identify the direct (victims and aggressors) and indirect (observers) involvement profiles, applying the “log-likelihood” option, which performs a probability distribution of the variables, to determine the distance or similarity between clusters. In particular, to identify the clusters or different profiles that we call “direct involvement profiles” (aggressors, victims) and “indirect involvement profiles” (observers), we used the answers to the questions: “Have you ever used violence against other students?” and, “Have you ever been the object of violence from other students?” And “Have you ever intervened when you saw someone using violence on other students?” for the indirect involvement clusters.

The “determine automatically” option was applied to determine the number of clusters. This procedure determines the “optimum” number of clusters automatically using Log-Likelihood, the criterion specified for clustering, in this case. This exploratory technique also provides data on the quality of the cluster, using the Silhouette Coefficient, which is a measure of the adequacy of cluster cohesion and separation, evaluating whether the structure resulting from forming the groups is: ≤0.25 insubstantial, 0.26–0.50 weak, 0.51–0.70 reasonable, o 0.71–1.00 strong.⁸² In this case, the mean silhouette was 1 (strong).

After classifying the cases, a MANOVA was performed to find between-group differences in the components of empathy (Perspective Taking and Empathic Concern). To estimate the effect size, the partial Eta squared was applied (ηp²), considering: <0.01 irrelevant, 0.01 small, 0.06 medium, and large from 0.14 on. To determine what profiles had significant differences, the post hoc Bonferroni comparison test was applied.

Results

Attitudes Toward Aggression, Perspective-Taking and Empathic Concern: Correlations

As shown in Table 1, favorable adolescent beliefs about the use of aggression (A_vs) correlated positively [r = 0.42, 95% CI (0.37, 0.46)], with a moderate association with the perception of strong support from parents for use of aggression in response to conflict (P_as). Similarly, stronger support for use of nonviolent strategies by adolescents (A_nvs) was positively related [r = 0.28, 95% CI (0.23, 0.33)] with the perception of strong support from parents for pacific solutions to conflict (P_nas), in this case, resulting in a weak association.

Table 1 Attitudes Toward Aggression, Perspective Taking and Empathic Concern. Pearson Correlation Matrix

The relationships established with both the cognitive (Perspective Taking, PT) and emotional (Empathic Concern, EC) components of empathy analyzed were negatively related to favorable adolescent attitudes toward aggression [PT: r = −0.28, 95% CI (−0.33, −0.23); EC: r = −0.30, 95% CI (−0.35, −0.24)], weak for Perspective Taking and moderate for Empathic Concern. Negative correlations were also found, although weak, with respect to the perception of parental support for aggressive solutions in response to conflict [PT: r = −0.19, 95% CI (−0.24, −0.13); EC: r = −0.14, 95% CI (−0.19, −0.08)].

Types of Involvement in School Violence

First, the following groups were identified by the possible combinations of direct involvement in school violence (Figure 1):

Figure 1 Direct involvement profiles (victims and aggressors). Abbreviations: M, mean; SD, standard deviation. Note: c1 = no direct involvement, c2 = victim-aggressor, c3 = victim, c4 = aggressor.

Cluster 1 (C1), labeled “no direct involvement”, was the most numerous of the groups and was made up of adolescents who said they had neither been assaulted nor used violence on other students. Cluster 2 (C2), labeled “victim-aggressor”, grouped those who responded affirmatively to both questions: “Have you ever been the object of violence from other students?” and “Have you ever used violence against other students?” Cluster 3 (C3), adolescents in the role of “victim”, was made up of those who said that they had been the object of violence from other students. Finally, Cluster 4 (C4) included adolescents who stated that they had used violence against other students, and this group had the role of “aggressor”.

Indirect involvement, based on the role of observer, was found by asking two questions: “Have you ever seen violence against other students?” which the subject only answered if he had witnessed an episode of school violence, and “Have you ever intervened when you saw someone using violence against other students?” where the answer was whether the participant had intervened or not.

The groups identified based on the participant answers to the two questions above were the following (Figure 2):

Figure 2 Indirect involvement profiles (Observers). Abbreviations: M, mean; SD, standard deviation. Notes: c1 = social desirability bias, c2 = active observer, c3 = non-observer, c4 = passive observer.

Cluster 1 (C1), where answers were incongruent, because the subjects stated that they had never witnessed episodes of school violence, and however, answered positively about their intervention in those cases. This group was characterized by the presence of incongruence or “social desirability bias”.

Cluster 2 (C2), the most numerous, was labeled “active observers”, and included adolescents who had intervened in episodes of school violence.

Cluster 3 (C3), labeled “non-observer”, included adolescents who said they had not witnessed episodes of school violence.

Finally, Cluster 4 (C4), called “passive observer”, included those who stated they had seen episodes of school violence, but had not intervened in them.

Differences in Empathy, by Involvement in School Violence

The MANOVA showed that there were significant differences between the direct involvement groups (Table 2) in the components of empathy (Wilk’s Λ = 0.963, F_{(6, 1222)} = 7.76, p < 0.001, η_p² = 0.19). Equality between the covariance matrices was examined using Box’s M test (M_Box = 9.80, F = 1.07, p = 0.377). The univariate analyses for each dependent variable revealed the existence of significant differences between the groups in Perspective Taking (F_{(3, 1222)} = 10.96, p < 0.001, η_p² = 0.26). Specifically, the group of “victims” scored significantly higher than the “no direct involvement” (p_bonf < 0.001), “victim-aggressor” (p_bonf < 0.01) and “aggressor” (p_bonf < 0.001) profiles. Between-group differences in Empathic concern were statistically significant (F(3, 1222) = 11.28, p < 0.001, ηp² = 0.27). In this case, statistically significant differences were observed in the “victims” group, who also had higher scores than the rest of the profiles: “no direct involvement” (p_bonf < 0.001), “victim-aggressor” (p_bonf < 0.01) and “aggressor” (p_bonf < 0.001).

Table 2 Direct Involvement Profiles and Empathy. Descriptive Statistics

Furthermore, the multivariate analysis applied to the indirect involvement groups (Table 3) revealed significant differences (Wilk’s Λ = 0.966, F_{(6, 1222)} = 7.20, p < 0.001, η_p² = 0.19). Equality between covariance matrices was examined using the Box’s M test (M_Box = 5.95, F = 0.66, p = 0.746). The univariate analyses found significant between-group differences in Perspective Taking (F_{(3, 1222)} = 4.53, p < 0.01, η_p² = 0.01), where the groups with the “social desirability bias” response and “active observer” scored highest. In particular, the “social desirability bias” group differed significantly from the “non-observer” (p_bonf < 0.05) and “passive observer” (p_bonf < 0.05) profiles. Similarly, the “active observer” group had scores significantly higher than the “non-observer” (p_bonf < 0.05) and “passive observer” (p_bonf < 0.05) profiles. In Empathic concern, there were also statistically significant between-group differences (F_{(3, 1222)} = 13.31, p < 0.001, η_p² = 0.03), where the highest mean score was in the group of “active observers”, showing statistically significant differences from the “non-observer” (p_bonf < 0.001) and “passive observer” (p_bonf < 0.001) profiles. The “social desirability bias” profile showed a significantly higher mean score than “passive observer” (p_bonf < 0.05).

Table 3 Indirect Involvement Profiles and Empathy. Descriptive Statistics

Discussion

The results confirmed the first hypothesis of this study, since it was observed that parental approval of the use of violent strategies promotes its normalization, and therefore, supports use of this type of behavior by adolescents.⁶⁰ While parental fostering of pacific solutions shows higher levels of empathy are related to less favorable attitudes of adolescents toward violence. Thus, parental promotion of strategies based on communication and socialization in the peer group,^47,48,50 along with better understanding and adherence to the emotions of other students was related to support of adolescents for non-violent solutions to resolve conflicts.^66–68

Furthermore, in regard to the second hypothesis of this study, the results on involvement in school violence analyzed showed the existence of different profiles within direct and indirect participation in school violence. Specifically, in direct involvement, four groups were found coinciding with what was suggested by Monteiro et al²⁰ youths with no direct involvement, victim-aggressors, victims and aggressors. The first was made up of adolescents who said they had not used violence or been the object of school violence was the most numerous, followed by victims, aggressors and victim-aggressors, respectively. Thus, the percentage of adolescents who admitted to having used violence against other students was lower than the percentage of victims who said they had been subjected to violence, as shown in the study by Fisher et al.²⁶

The clusters of the indirect involvement profiles, as observed, included a high percentage of active observers, characterized by helping victims of school violence.²⁹ This was followed by the passive observers, who did not react to situations of abuse of other students due to their fear of the repercussions.^30,32 And, finally, the lowest percentage was of “non-observers”, who said they had not witnessed episodes of school violence. In addition, a group with incongruent responses was observed, in which, motivated by social desirability, adolescents said they had never been witnesses of violence, but they had, however, intervened in defense of the victim.

Once the different groups of involvement in school violence had been defined, the results within the direct involvement groups showed the significantly highest scores in the group of victims on Perspective Taking and Empathic Concern. Likewise, after victims, those in the “no direct involvement” group had the highest scores on the two dimensions of empathy evaluated. This seems to show that an empathic attitude promotes interpersonal relationships, lessening the use of violent behavior.⁹

Finally, in the multivariate analysis applied to the indirect involvement groups, the results showed a direct relationship between the components of empathy and active involvement of the observer, both in Perspective Taking and Empathic Concern. Active observers are able to put themselves in the place of the victims and feel concern for them, and then take action to help them in situations of violence. Similar results were shown in the studies by Evans et al²⁹ and Bauman et al³³ where observers who intervened when other students were subjected to violence had high scores in empathic response to conflictive situations. In view of all of the above, it is fundamental to encourage adolescents to intervene in situations of school violence.

Based on the empirical results of this study, along with the theoretical review, we can say that both transmission of parental values and individual characteristics such as empathy, have an important role in the formation of favorable or unfavorable beliefs and attitudes on the use of violent strategies for solving interpersonal conflicts in adolescence. Empathy was shown to be another important variable in the involvement of youths in school violence. Thus, showing more empathy was related to not using violence against other students and defending those who are the object of such violence.

Limitations

This study had some limitations which should be borne in mind. In the first place, it should be emphasized that the responses on the point of view of the parents about violent behavior were provided by the adolescents themselves, and therefore, these results should be compared to parents’ opinions. In the second place, the study focused on two components of empathy, Perspective Taking and Empathic Concern, which could lead to bias if the results are generalized to the rest of the components of empathy. The difficulty for generalizing these findings outside of Spain or other geographic regions should also be mentioned, as the profiles for involvement in violence may not coincide with youths in other areas.

The results of this study suggest that in future research, the sample size be increased and preuniversity [bachillerato] students added to the sample. It would also be of interest in future work to include other variables, such as emotional intelligence and self-esteem in the analysis as factors related to violent behavior.

Strategies for intervention for reducing violent behavior in schools could be designed based on the results of this study.

Conclusion

This study inquired into the role that parental perspective and empathy have in support for violence by youths, and also the levels of involvement in school violence. The results showed that parental support for the use of nonviolent strategies for resolving conflicts, as well as higher levels of cognitive and affective empathy, are related to more favorable adolescent beliefs in pacific solutions. Students who did not use violence on their classmates, and who acted when in the presence of bullying of another student had higher levels of empathy.

There is therefore a need to train parents in alternative nonviolent conflict resolution techniques so that the education youths receive at home is based on peaceful socialization. Moreover, school violence prevention programs based on training in individual variables, such as empathy, could be options promoting identification of students with their peer group. Thus, violence toward other students would be reduced and action by those who witness this type of action encouraged, creating a network of support and protection of victims of school violence.