Cyberbullying Involvement: Impacts of Violence Exposure in The Media, Family, Society, and School

Perundungan maya adalah masalah serius yang dapat terjadi pada remaja ketika berkomunikasi melalui internet. Penelitian ini menginvestigasi paparan kekerasan pada beberapa konteks seperti pada media, keluarga, masyarakat dan sekolah terhadap kaitannya pada keterlibatan siswa melakukan ataupun menjadi korban perundungan maya. Melalui survei kepada siswa menengah atas (total N = 201), studi ini menemukan bahwa paparan kekerasan, terutama pada media dan sekolah, terbukti secara signifikan membedakan status keterlibatan siswa dalam perundungan maya. Remaja yang terpapar kekerasan pada media dan sekolah, cenderung menjadi pelaku sekaligus korban perundungan maya. Analisis tambahan pada jenis kelamin partisipan menunjukkan adanya perbedaan yang signifikan dimana perempuan lebih cenderung menjadi korban cyberbullying, sedangkan laki-laki cenderung menjadi pelaku cyberbullying. Selanjutnya, implikasi mengenai pencegahan perundungan dibahas dalam penelitian ini.


Introduction
Due to the recent communication technology evolution, the number of cyberbullying increases because the bullying is not only committed faceto-face directly but also mediated through the internet. This cyberbullying issue has become a concern and attention in many countries. In Germany, 5.4% of students admitted becoming cyberbullying victims for at least once in a week or more (Riebel, Jager & Fischer, 2009). Meanwhile, research with Serbian adolescents as the samples reported that adolescents as the cyberbullying perpetrator were approximately 10%, and cyberbullying victims were around 20% (Popović-Ćitić et al., 2011). Canada had about 4.5% to 33.7% of cyberbullying prevalence cases on adolescents (Riddell, Pepler, & Craig, 2018). In the meantime, an Asian country, i.e., South Korea, was predicted to have approximately 34% of teenagers involved in cyberbullying cases (Lee & Shin, 2017). Indonesia is also susceptible to cyberbullying as other countries. A study mentioned that almost 80% of Indonesian students admitted being a cyberbullying victim (Safaria, 2016).
Contrary to face-to-face bullying, cyberbullying victims cannot elude because they do not have any places to hide or transfer to other schools or move their house. The characteristics of internet, which enable persons to keep in touch for 24/7 and at anytime and anywhere, and also the anonymity of the internet which allows the perpetrator to hide, make it possible for cyberbullying to have bad risks for the victims twofold rather than the face-to-face or direct bullying (Ovejero et al., 2016). Various researches have mentioned adverse effects of the cyberbullying, namely loss of confidence (Patchin & Hinduja, 2010), depression (McDermott, 2011), student's declining academic performance (Faryadi, 2011), and even suicide tendency (Patchin & Hinduja, 2010). A physical absence of the victims before the bully also caused the lack of guilty feeling and might trigger the bullying to be more aggressive (Ovejero et al., 2016). Willard (2007) mentioned several types of cyberbullying: (1) flaming (send a rude message about someone), (2) online harassment (repeatedly send degrading, offensive, and hurtful words via email and a short text message), (3) denigration (send detrimental, fallacious, or rumour messages about someone to other persons or post such messages online), (4) impersonation (masquerade as someone else by hacking the victim's account, and send or post material which makes the victim look bad, put her/him in a trouble or danger, or damage someone's reputation or friendship, (5) outing and trickery (send or post material which contains shameful, sensitive, and personal information, including forwarding a private message or picture by tricking someone to reveal her/his shameful secret or information. Then, such message or image will be disseminated online), (6) exclusion (deliberately exclude or expel someone from an online group), (7) cyberstalking (repeatedly send a threating message or other online activities, which cause someone feels afraid for her/his personal safety).
Many experts stated that the cyberbullying was closely related to the impacts of violence exposure (Tzani-Pepelasi et al., 2017;Lam et al., 2013;Calvete et al., 2010). Violence exposure shows anti-social behavior. Thus, imitation mediated by the internet is possible. However, such a relationship has not been investigated. Based on such the background, a research question of this study was formulated as follows: How is the description of the violence exposure from several aspects, such as violence in media, family, soci-ety, and school, independently related to a cyberbullying involvement tendency? Moreover, this research is also aimed to explore the background of adolescents involved in the cyberbullying. The background exploration is expected to improve our understanding of the cyberbullying behavior. Ovejero et al. (2016) define bullying as an aggressive action to hurt someone both verbally and physically. This aggressive action leads to physical and verbal violence. This bullying is divided into two categories, namely traditional bullying and cyberbullying. The basic principle of conventional bullying and cyberbullying is the same, which is aimed at hurting someone. But between the two also have differences. The first difference is that conventional or traditional bullying is direct or face to face while cyberbullying is bullying indirectly using electronic communication technologies such as e-mail, cell phones, text messages, and chats. The second difference is that conventional bullying tends to contain physical violence, while cyberbullying contains more verbal violence.

Literature Review
From several previous studies, it can be assumed that there are several situational variables that influence bullying, namely: First, ownership of power is higher (Ovejero et al., 2016). When someone feels they have more ability than others, then an exclusive feeling, discomfort, dislike, hatred towards the presence of other people who have different identities with them will emerge.
Second, the supervision of teachers and parents is low (Swearer & Doll, 2008). Bullying usually occurs during the least supervised time. This indicates that when there is no or weak supervision by teachers and parents, that is where bullying will occur.
Third, the level of relationships closeness in family and friends. Someone who has a good relationship in family and friends usually has low psychological distress (Horwitz, Reynolds & Charles, 2015). Bebanic et al. (2017) show that psychological distress and exposure to violence has correlated.
The three situational variables above are derived from personal, which is more often found within individuals as well as interactions between individuals. Other situational variables are situational variables sourced from media exposure (non-personal). Based on Bushman & Huesmann (2006), when children or adults exposed to violent content from the media, then they tend to involve aggressive behavior. The short-term effects of violent media were greater for adults, whereas the long-term effects were greater for children.
Besides exposure from media violence exposure, social environment may have an important role. When the high intensity of violence exposes someone in their social environment or in neighbor relationships, then the person may have experienced negative emotional consequences (Joshi & Kaschak, 1998). The experience of negative emo-tions can be a precursor of the severe psychological problems in adolescents (Larson & Asmussen, 1991), one of which is juvenile delinquency.
From the explanation above, overall, this study would like to describe the violence exposure in the family, school, society, and the media towards cyberbullying behavior in adolescents.

Respondents
The research questionnaires were distributed to 236 students of twelve senior high schools in Jakarta area. The respondents are participants of a campus visit or an event to introduce environment of a university for high school students. They were asked to participate in research about the cyberbullying on teenagers voluntarily. The respondents were informed that this study was anonymous and did not have any impacts on participating individuals. Thus, it was expected that they could answer the questions honestly as their experience. During the research, the participants were asked to fill out independently the provided questionnaire (self-reported questionnaire) for 15 minutes. Out of 236 students, only 201 participants answered all questions in this study and could be analyzed further.

Respondents' Background
The questions used in this study were taken from various literature under several considerations. For example, to know the respondents' Social Economic Status (SES), this study chose the Family Affluence Scale (FAS). FAS was chosen to measure the socioeconomic status because, according to Currie et al. (1997), the SES measurement in children and adolescents were often obstructed by children or adolescents' accuracy in reporting their parents' income. Meanwhile, in FAS, the SES is indirectly measured by asking four points on the respondents' family, such as possession of cars, private bedroom, vacation frequency, and possession of the computer.
FAS has four questions which scores are weighted, i.e., first, "Does your family own a car?' Answer No (Score: 0), Yes, one car (score:1), Yes, two cars or more (Score: 2); second, "Do you have a personal bedroom, which you do not share with anybody else?" Answer No, I share a room with another family member (Score: 0), Yes, I have a bedroom only for me (Score: 1); third, "In last year, how many times did you go to vacation with your family?" The answer never (Score: 0), Once (Score: 1), Twice (Score: 2), More than twice (Score: 3); fourth, "How many computers or laptop does your family own?" Answer None (Score: 0), One (Score: 1), Two (Score: 2), and more than two (Score: 3). This study divided FAS score calculation result into three categories, namely: low, middle, and high SES. Each participant's answers determined the categorization. Total FAS answer scores of 0-2 were grouped into low SES, 3-5 as middle SES, and 6-9 as high SES.
Furthermore, to have further information on the respondents' background, the questionnaire also asked about parental supervision, internet use duration, average academic performance, parental status, and face-to-face/direct bullying. The face-to-face/direct bullying was measured by asking whether the respondents have experienced any verbal/physical bullying by their school peer/ junior/senior. The response choices are binary "yes" and "no". Then, the responses were grouped into four categories, i.e., perpetrator-victim; perpetrator; victim; non-perpetrator and victim.

Violence Exposure in the Media
Questions on the violence exposure in the media were modified from a study carried out by Joshi & Kaschak (1998). The questions asked the respondents to choose an answer from "never" to "always" for a violence exposure frequency in TV shows or films (example: "have you ever watched a person shooting somebody else in a TV show or film?"). The question used the Likert Scale, and Cronbach's alpha was 0.69.

Violence Exposure in Schools
This study took three questions about an experience to see violence in schools, and they were adapted from the study conducted by Flannery et al. (2004). Three specific violence types were asked, i.e., threats, assault with a weapon, and assault without a weapon. The Likert scale, which had five choices from "never" to "always" was applied to observe the frequency of each violence type observed in school. The Cronbach's alpha score obtained was 0.63.

Violence Exposure in Family
Four questions investigated the violence exposure in the family. The items included the frequency of adults living in the house having fights and arguments to cause conflicts, assault in the home, and one of the family member feel threatened in the home. The Cronbach's alpha score in this scale was 0.61.

Violence Exposure in Neighbourhood (Community/Society)
Questions on violence exposure in the community near the house were adapted from several studies, i.e., studies by Koposov & Ruchkin (2011), Joshi & Kaschak (1998, and Dubow et al. (2010). There were eight questions which asked the frequency from "never" to "always" regarding an experience to encounter violence in the community near the respondents' house, such as a person be-ing assaulted, robbed, shot, or injured by a weapon. The Cronbach's alpha score in this scale was 0.63.

Cyberbullying
The cyberbullying was measured using the European Cyberbullying Intervention Project Questionnaire (ECIPQ). The validity has been tested in several countries (Del Rey et al., 2015;Herrera-López et al., 2017). ECIPQ was differentiated into two question categories, namely cyber victimization and cyber aggression. The questions used a scale from "never" to "always." Eleven questions on cyber victimization were asked. Examples of the questions are "someone threatens me through social media/internet," and "someone says something bad about me to other persons through internet/social media." The Cronbach's alpha score for this question group was 0.73. Next question category is questions on cyber aggression. Examples of these questions are "I viciously kick out someone from a chatting group/social media," and "I post someone else's embarrassing video or picture on the internet/ social media." The Cronbach's alpha score for this category was 0.76.
For the analytical purpose, collected data would be processed by grouping these ECIPQ results into four categories, i.e., cyberbullying perpetrator-victim category if the respondents chose other than "never" for at least one question on cyber-victimization and cyberaggression. Cyberbullying perpetrator category if the respondents answered other than "never" for at least one question on cyber-aggression, but responded "never" for cyber-victimization questions. The next category is cyberbullying victim, if the respondents answered other than "never" for at least one question on cyber victimization, but responded "never" for cy-beraggression questions.
The non-cyberbullying perpetrator and non-cyberbullying victim category if the respondents selected "never" on cyberaggression and cyber-victimization.

Data Analysis
The descriptive and inferential statistics were conducted for data analysis. The descriptive statistics were used to identify the participants' background information, such as gender, age, socio-economic status, daily internet use, parental status, and academic performance in schools. The Chi-square and ANOVA was selected to test whether there was a significant variance on the bullying victim and perpetrator status related to various violence exposures, direct bullying experiences, and participants' background.

Descriptive Statistics
Out of a total of 201 respondents, 114 (56.7 percent) respondents were female, and 87 (43.3 percent) were male. The respondents' age ranged between 15-18 years old, where four persons (2 percent As referred to in table 1, most respondents come from the upper socio-economic background. Only 6 percent of the respondents have the low socio-economic background, and 35.8 percent come from the middle socio-economic background. Then, there was an interesting finding, i.e., 88 percent of the respondents admitted that their parents never supervised them when they used the internet. The remaining 12 percent responded that they were supervised by their parents. Meanwhile, almost half of the participants (45.7 percent) used the internet for more than 7 hours every day, and 34.8 percent used the internet for 4-6 hours daily.
Most participants responded that their academic performance was equal to the class average. Then, the majority of the respondents (79.1 percent) had a dual parent (father and mother). According to such data finding, it could be stated that 80.5 percent of the respondents, who had dual parent, used the internet for more than 4 hours every day without any supervision from their parents.
Related to the bullying experience, most respondents (65.2 percent) answered that they were neither a perpetrator nor victim of a face-toface or direct bullying. 17.9 percent of the respondents were classified as the perpetrator of faceto-face or direct bullying. Then, 7.5 percent were a victim, and 19 percent were perpetrator-victim in face-to-face or direct bullying.
This finding was contrary to the cyberbullying experience, where almost half of the respondents (45 percent) were grouped into the cyberbullying perpetrator-victim. It is consistent with the study carried out by Deniz (2015), which stated that the higher the socio-economic status, the higher the tendency for participating in cyberbullying. Likewise, a study conducted by Tippett and Wolke (2014) showed that the lower the socio-economic status, the higher the tendency to be involved in direct bullying as the perpetrator-victim; even though the correlation was not highly significant.

Characteristics of the Cyberbullying Perpetrator and Victim Status
When the cyberbullying involvement status was separately observed in each research variable, several statistically significant differences could be found in some variables, i.e., gender, violence exposure in schools, violence exposure in the media, and face-to-face/direct bullying experience. Table 2 provides a summary of the detailed comparison between the cyberbullying perpetrator and victim, which is related to the demographical background, parental supervision, internet use, violence exposure, and face-toface/direct bullying experience.
There was a significant difference between the cyberbullying perpetrator and victim's gender and status (Χ 2 = 8.606, p = 0.035). In this finding, male was more inclined to be a cyberbullying perpetrator (59 percent), while female was more likely to be a cyberbullying victim (71 percent), regardless the number of male and female, who was the perpetrator-victim, was relatively equal. This finding is consistent with several studies by Li (2006), Festl and Quandt (2013), Navarro and Jasinski (2013), Laftman et al. (2013), Heiman et al. (2015, Schultze-Krumbholz et al. (2015), Wachs et al. (2015). Nevertheless, many other kinds of research stated a different result for the correlation between gender and cyberbullying.
In general, male is always associated with direct aggressive behavior than female, such as a physical fight (Lansford et al., 2012). It corresponds with the face-to-face/direct bullying characteristics. Cyberbullying is an indirect aggressive behavior because it does not have any physical contacts. Thus, it should have a significant correlation with female.
However, this study had a different finding, where female was more inclined to be the cyberbullying victim, and male was the perpetrator. A possible interpretation of such finding is different masculinity and femininity distinguishing the cyberbullying perpetrator and victim, instead of different gender.
Related to the violence exposure, there were only two violence exposures, which significantly differentiated the cyberbullying perpetrator and victim, namely violence exposure in schools (F = 6.028, p < 0.01) and violence exposure in the media (F = 2.750, p = 0.044). From this finding, either the cyberbullying perpetrator or victim had a higher average score for the violence exposure in schools and the media than other categories. This finding should be considered by parents and schools to create a condition where the students will not expose to the violence in school. Furthermore, parents may provide support by selecting media which have safe contents for children or adolescents.
In addition, table 2 also shows a significant difference in face-to-face/direct bullying experience for four cyberbullying perpetrator and victim status categories (Χ 2 = 28.36, p < 0.01). Forty-nine percent of the respondents, who fell into the cyberbullying perpetrator-victim category, admitted that they were neither face-to-face/direct bullying perpetrator nor victim, and 27.8 percent of the respondents in such category was involved as the face-to-face/direct bullying perpetrator. 24.1 percent of the cyberbullying perpetrators were also face-to-face/direct bullying perpetrators. This finding is consistent with the study by Erdur-Barker (2010), which stated that traditional face-to-face bullying was correlated to the cyberbullying. However, it should be noted that the correlation between cyberbullying and direct bullying was not highly significant (Ybarra et al., 2007). However, there are limitations to this study. First is the less varied socioeconomic status of the respondents. In this study, majority of the participants came from high socio-economic status (SES) background. Previous studies explained that SES was correlated to cyberbullying involvement. Therefore, future researches may try to explore the violence exposure impacts related to cyberbullying and SES as a control variable.
Second, violence exposure in media is only related to violence exposure through TV shows and films in this study. Future studies may attempt to extend the violence exposure in media via games, internet, and music which has not been observed in this study. Third, this study has a statistical limitation because the sample size is relatively small. Future studies could be conduct-ed by increasing the number of participants and better sampling method design for generating generalizability.

Conclusion
This research attempts to contribute to the studies of violence exposure impacts to cyberbullying involvement by adolescents in several settings. The effects of violence exposure in the media and schools are proven to correlate with the cyberbullying involvement by teenagers. Parents and schools should pay attention to this finding, where parents must select media which do not show violence and schools should consider some policies, such as an anti-bullying program, and create a safe school environment.