“If it’s really excessive, it can enter your heart”: A Mixed Methods Investigation of Bullying Among Early Adolescents in Semarang, Indonesia

Bullying is a major public health concern for Indonesian adolescents, with well-documented links to a range of emotional and behavioral problems. Despite such evidence, few investigations have employed qualitative methods to illuminate youth’s own perceptions of bullying and its psychosocial correlates in this context. The current study aimed to address this gap through an exploration of Indonesian adolescents’ motivations, perceptions, and beliefs regarding bullying. Building on prior quantitative findings, an explanatory sequential mixed methods approach was used to better understand the myriad ways in which bullying ties into other psychosocial challenges. Qualitative interviews were conducted with a total of 45 adolescents ages 13 to 14 (25 girls and 20 boys) in two junior high schools in Semarang between October and December 2019. Interviews were audio-recorded, transcribed verbatim, and translated into English for analysis. Qualitative data were then coded using an inductive thematic analysis approach. Interviews yielded contextual insights into adolescents’ definitions of bullying including the distinction between “normal” and “serious” bullying; related risk behaviors; key drivers; social and emotional consequences; and coping strategies. Across these thematic categories, a number of noteworthy gender differences emerged, highlighting the role that underlying gender norms can play in driving bullying involvement. Further, findings emphasize the need to develop a locally valid definition of bullying which takes into account the ways in which emotional distress may be both a criterion and a consequence of bullying. Findings can be used to inform bullying prevention programs targeting Indonesian youth.


Introduction
Bullying is one of the most common forms of youth violence worldwide and is a significant public health challenge among adolescents across diverse country settings. Cross-national studies have suggested that approximately 25% to 35% of adolescents globally are involved in bullying, although country-level prevalence estimates are highly variable (Craig et al., 2009;Modecki et al., 2014). Bullying has traditionally been defined as repeated aggressive behaviors that are intended to cause harm and involve a power imbalance between the perpetrator and the victim (Olweus, 1993). Such behaviors may include acts of physical aggression (e.g., hitting, pushing), verbal aggression (e.g., namecalling, threatening), relational aggression (e.g., social exclusion), and increasingly, cyber aggression (e.g., online harassment) (Menesini & Salmivalli, 2017). While previous research tended to characterize youth as either perpetrators or victims of bullying, the field has increasingly recognized that such labels are not mutually exclusive, with many youth falling into both categories (i.e., "bully-victims") (Cook et al., 2010).
A large body of research has documented the bidirectional relationships between bullying and other psychosocial challenges. Adolescents involved in bullying as victims, perpetrators, or both are at increased risk of a range of short-and long-term emotional and behavioral problems, including depression, anxiety, suicidal behaviors, aggression, and substance use (Moore et al., 2017;Zych et al., 2015). Likewise, longitudinal investigations examining the psychosocial antecedents of bullying involvement have identified internalizing symptoms, conduct problems, and social challenges as important predictors (Kljakovic & Hunt, 2016). Notably, studies have found that bully-victims may be uniquely vulnerable, with higher rates of mental health and psychosocial problems compared to those involved in bullying purely as victims or perpetrators (Cook et al., 2010;Copeland et al., 2013).
While the extant literature speaks to the pervasive nature and detrimental impacts of bullying involvement, it has notable limitations. In particular, existing research can be difficult to interpret due to considerable inconsistencies in the measurement of bullying (Vivolo-Kantor et al., 2014). Some of these inconsistencies relate to the lack of a globally accepted standard definition of bullying. For instance, while definitions of bullying frequently specify that it must be repeated (e.g., Olweus, 1993), there are differing opinions regarding whether repetition is required given that a single incident can have harmful effects (Swearer et al., 2010). Likewise, despite widespread recognition of the growing burden of cyberbullying, experts disagree on whether cyberbullying is phenomenologically distinct from "traditional" forms of bullying (Olweus & Limber, 2018). Studies have also found that different measurement strategies-in particular, providing youth with a formal definition of bullying compared to simply describing bullying behaviors-can have a drastic impact on reported bullying prevalence (Green et al., 2013;Vaillancourt et al., 2008). Finally, cross-national research has suggested that there are important nuances in the ways that adolescents from diverse cultural contexts understand, define, and describe bullying (Murray-Harvey et al., 2010).
In addressing these issues, researchers have called for an increase in studies that utilize qualitative methods to illuminate youth's own perceptions of bullying (Hutson, 2018;D. U. Patton et al., 2017). Such studies complement quantitative research by allowing for a deeper and more contextual understanding of bullying involvement and can ultimately be used to inform appropriate assessment tools and intervention approaches (D. U. Patton et al., 2017). Indeed, existing studies that have adopted qualitative methodologies have uncovered important inconsistencies between researchers' definitions of bullying compared to how adolescents themselves conceptualize this phenomenon. A common finding across studies is that while youth often describe aggressive behaviors as bullying, they frequently omit the criteria of intentionality, repetition, and power imbalance from their definitions (Guerin & Hennessy, 2002;Hellström et al., 2015;Vaillancourt et al., 2008).
In Indonesia, the setting of the current study, existing research suggests that bullying is a prevalent concern for adolescents. According to nationally representative data, around 20% of youth report past-month experiences of bullying victimization, placing them at heightened risk of loneliness, social isolation, and substance use (Yusuf et al., 2019). Such findings have driven the Indonesian government to adopt several initiatives aimed at reducing interpersonal violence in schools, such as the Child Friendly Schools Initiative and the National Strategy for the Elimination of Violence against Children (Arifin et al., 2019;Ministry for Women's Empowerment and Child Rights, 2015). To date, however, there have been few qualitative studies that attempt to uncover perceptions of Indonesian adolescents regarding bullying and its psychosocial correlates. Given the paucity of evidence-based interventions targeting bullying involvement in Indonesia (Bowes et al., 2019), such information could be transformative in shaping locally acceptable psychosocial support strategies for vulnerable youth.
The current study attempts to fill these gaps through a mixed methods exploration of bullying involvement among Indonesian youth. It was nested within the Global Early Adolescent Study (GEAS), a longitudinal investigation of risk and protective factors for healthy development among early adolescents (ages 10-14) living in low-resource urban settings worldwide. The current study employed an explanatory sequential mixed methods design, in which qualitative data were collected after an initial quantitative phase and used to interpret and expand upon quantitative results (Creswell & Clark, 2017). Specifically, this study followed up on quantitative findings from a prior analysis of co-occurring emotional and behavioral problems among GEAS adolescents in Indonesia (Fine et al., 2022) with qualitative interviews intended to yield a more contextualized understanding of the contribution of bullying involvement to an adolescent's broader constellation of psychosocial risks. Qualitative interviews were conducted with a subsample of GEAS participants living in Semarang. While a range of issues were covered, the current analysis focuses on (a) adolescents' definitions of bullying, (b) their perceptions of the causes and consequences of bullying, and (c) their strategies for coping with bullying experiences.

Study Setting
In Indonesia, the GEAS includes three study sites: Semarang, Java; Bandar Lampung, Sumatra; and Denpasar, Bali. As outlined above, this explanatory sequential mixed methods study expands upon preliminary quantitative findings from across the three Indonesian sites through qualitative data from a subsample of adolescents living in Semarang. The capital of the Central Java province, Semarang, was selected as the site for qualitative data collection for several reasons. While the city has enjoyed notable economic development over the past several decades, infrastructure strain driven by unchecked population growth has contributed to increasing marginalization among the urban poor (Giyarsih & Marfai, 2017;Semarang City Government, 2016). The potential for such urbanization patterns to increase psychosocial vulnerability (Lu, 2010) makes Semarang a particularly important environment in which to study behavioral risks among disadvantaged youth. Further, there is indication that the prevalence of mental disorders in Central Java may be higher than other regions of the country (Indonesia Agency of Health and Development, 2013) stressing the need for investigations that explore antecedents of mental health problems in this context. Qualitative data were collected in Semarang between October and December 2019.

Participants and Procedures
Sampling for the qualitative study was based on quantitative findings from a latent class analysis (LCA) of GEAS adolescents from across Indonesia (N = 4,657). Detailed methodology and results from this quantitative analysis are described in a companion article (Fine et al., 2022). Briefly, LCA was used to identify subgroups (i.e., classes) of adolescents based on their patterns of endorsement to a set of 10 emotional and behavioral indicators. These included five dichotomous indicators related to symptoms of depression and anxiety, two related to aggressive behavior, two related to peer victimization, and one related to substance use ( Table 1). The class enumeration process involved testing a series of models with an increasing number of classes, with fit indices compared to determine the best-fitting model (Collins & Lanza, 2010). In addition to fit indices, the theoretical interpretability of classes was considered. Results from this process supported a four-class solution in Indonesia, which consisted of the following subgroups: Well-Adjusted (49%), with few emotional and behavioral problems; Emotional Problems (29%), with elevated symptoms of depression and anxiety; Behavioral Problems (15%),with increased involvement in aggressive behaviors, peer victimization, and substance use; and Maladjusted (6%), with cooccurring emotional and behavioral problems (Figure 1).
Following the class enumeration process, qualitative study participants in Semarang were purposively sampled from each of the four latent classes. While we originally intended to make qualitative comparisons between adolescents from different classes, variable data richness across interviews made it infeasible to explore such thematic differences. Instead, we have used this stratified purposive sampling strategy (M. Q. Patton, 1990) to ensure representation from adolescents with varied patterns of emotional and behavioral problems. In order to arrive at our sample size goal of 40 adolescents, around 16 participants were selected from each latent class, equally divided between Age assumes interviews conducted 1 year after quantitative baseline assessment. b Coded so that positive endorsement indicates "agree a lot" or "agree a little." boys and girls. While the Indonesian GEAS included adolescents who were 10 to 14 years old at baseline, sampling was limited to those who were 12 to 14 years old as they were expected to be able to provide richer qualitative information. As interviews were conducted approximately 1 year after quantitative data collection, however, participants were 13 to 15 years old at the time of the qualitative study.
Of the 63 adolescents who were selected from two participating junior high schools, 45 consented to participate (Table 1). This included 11 participants from the Well-Adjusted class (six boys and five girls), 9 from the Emotional Problems class (four boys and five girls), 12 from the Behavioral Problems class (six boys and six girls), and 13 from the Maladjusted class (four boys and nine girls). Given the salience of bullying in adolescence, this sample size was deemed sufficient for achieving theoretical saturation (Sandelowski, 1995). Interviews explored adolescents' motivations, perceptions, and beliefs regarding bullying involvement. They were implemented using a semi-structured interview guide, which was developed in English and subsequently translated and back-translated by local translators. The interview guide was piloted with nine adolescents to ensure that questions were acceptable, comprehensible, and age-appropriate. Interviews took place in a private room at school and lasted between 45 and 90 minutes. They were conducted in Bahasa Indonesia by three experienced Indonesian researchers. Prior to data collection, the interview team participated in a 3-day training, which included sessions on the research protocol, qualitative interviewing methods, and human subjects protection. Interviews were audio-recorded and transcribed verbatim, and field notes were written immediately after each interview. The interview team participated in daily debriefing sessions in order to address challenges that arose, discuss emergent themes, and plan for future interviews. Ethical approval was obtained from the Institutional Review Board of the Johns Hopkins Bloomberg School of Public Health and the Medical and Health Research Ethics Committee of the University of Gadjah Mada. Prior to data collection, informed consent was obtained from parents and assent was obtained from adolescent participants.

Data Analysis
Transcribed audio-recordings and typed field notes were translated into English on an ongoing basis, with translated data made immediately available to facilitate iterative analysis. These qualitative data were coded using an inductive thematic analysis approach (Braun & Clarke, 2006) that drew on elements from grounded theory (Charmaz, 2006). Specifically, analysis began with a process of open coding, whereby lines of data were examined and labeled based on the thematic content within. In order to ensure that codes were grounded in the original data, the first phase of open coding was carried out collaboratively by the lead author alongside the interview team, with the lead author coding in English and the interview team coding in Bahasa Indonesia. After coding three representative transcripts, initial codes were compared, grouped into tentative categories, and organized into a preliminary bilingual codebook. To refine the codebook, a further set of 10 transcripts were then coded independently by the lead author.
Following open coding, the lead author applied the codebook to the full set of qualitative data, including translated transcripts and field notes. Axial coding was used to explore relationships within and between categories, develop emergent themes, and draw substantive conclusions supported by the data. Matrices were also employed to examine overlaps between key codes and differences between boys and girls. MAXQDA, a qualitative data analysis software program, was used to help organize and analyze the data (VERBI Software, 2019). Throughout data analysis, memos were used to document the analytic process. The lead author also discussed findings with the interview team on an ongoing basis to ensure agreement on emergent themes.
In order to increase the credibility of results (Lincoln & Guba, 1985), dissemination meetings were held at each junior high school to present preliminary findings and obtain feedback from key school personnel on points of agreement, points of disagreement, and missing information. These meetings were facilitated by the interview team and included school principals, counselors, and teaching staff. Finally, following the qualitative data analysis process, qualitative conclusions were integrated with quantitative results in order to consider important points of convergence and dissonance (Onwuegbuzie & Teddlie, 2003).

Types of Bullying
Participants had a range of perspectives on what constituted bullying. Nearly all of the adolescents described situations of verbal bullying, which most commonly included mocking or insulting others about their family background, physical appearance, or perceived shortcomings. Relatedly, a number of adolescents mentioned name-calling (e.g., "Fatty," "Blacky," "Slut"), which was generally carried out over a prolonged period of time. According to participants, a particularly ubiquitous form of verbal bullying was the use of an individual's parent's name or profession as a form of taunting, as detailed below: So my father is of Chinese descent . . . My father opened a wholesale store which supplied a lot of different markets in my area. The kids called me "Chinese Market" or "Chinese-Javanese." (Girl,age 13) Many adolescents also mentioned physical bullying, such as hitting, kicking, shoving, pinching, spitting, and hair-pulling. Physical bullying was thought to be much more common among boys, although several adolescents mentioned that girls sometimes participate in less severe forms (e.g., pinching or hair-pulling). By contrast, relational bullying was described as largely unique to girls, and included spreading rumors, socially excluding, and ostracizing. The dynamics of relational bullying were also distinct from verbal and physical bullying: whereas verbal and physical bullying were seen as being either one-on-one or group-on-one, relational bullying was nearly always characterized as being a group ganging up on a single victim. Both physical and relational bullying were depicted as being frequently accompanied by verbal bullying, as in the following examples: There was this one time when my friend had brought lunch with her and was about to eat, but the others bullied her, made fun of her food until my friend couldn't even eat. I was about to say sorry but I was afraid to do it . . . Yeah, [during that time] they stayed away [from her]. Continued to make fun of her, and talked behind her back . . . Anyway, she was completely alone. No one befriended her. (Girl, age 13) As for me, maybe . . . in his eyes, I am weak. That's why he punched me . . . He hit me and insulted me, usually. (Boy, age 13) While less commonly mentioned, several participants brought up cyberbullying. In particular, participants recounted instances in which adolescents would mock or harass others either through the comments section on social media platforms or through social media status updates. Finally, a number of adolescents labeled sexual harassment as a type of bullying, including groping and making sexual comments. While sexual harassment was largely described as boys targeting girls, a few participants discussed boys who harassed other boys, such as the adolescent below: He mocks the boys, too. And he grabs their butts and genitals . . . there was one time where a boy was writing something down. The seat next to him was empty, so he sat next to him. He touched his genitals, and then continued to bully him later. He'd say things like, "You're such a sissy, you play like a girl." Or he'd say, "Sissy, you don't have a penis and you have flat butt." (Girl, age 14)

Serious Bullying
In defining bullying, a key theme across many of the interviews was the distinction between "serious" and "normal" bullying. Many of the behaviors described by participants were seen as part of the normative dynamics between adolescents. This was particularly true of mocking, which was depicted as a "common" way for adolescents to "joke" or "play around" with each other. While participants generally agreed that bullying could sometimes be serious, they held differing opinions regarding factors that distinguished serious from normal bullying. For some, serious bullying was related to the intent of the perpetrator: it is serious if someone intends to cause harm or "means what they say." For others, serious bullying was related to the consequences for the victim: it is serious if someone "doesn't like it," "takes it personally," or ends up with "hurt feelings." The frequency of bullying also contributed to its seriousness, with bullying that was "continuous" or "everyday" considered to be more severe by several participants. Further, many adolescents noted that mocking which started as a joke could "go too far" if it was excessive, thereby "crossing the line" into bullying.
If they make fun of me once or twice, they're probably just joking, it's normal, but if they keep mocking me, then I'll consider them to really mean it. That kind of mocking can enter your heart. Usually mocking doesn't enter your heart, but if it's really excessive it can enter your heart and can even cause sakit hati [sick heart]. (Girl,age 14) Several adolescents also suggested that different types of bullying were more or less serious, with physical bullying generally considered to be more severe than verbal bullying. By contrast, a few participants felt that verbal bullying was just as serious because it attacked someone's "pride" or "good name," and thus necessitated a response. Finally, a number of participants indicated that boys generally bully harder than girls, both because they are more likely to "go too far" with their mocking and because they are prone to perpetrate physical violence.
If they don't get physical, it's fine. You're not supposed to do it physically. (Boy, age 12) Boys tend to [bully] more seriously, and they also tend to get into more fights . . . [boys] tend to do it more, so they often cross the line. (Girl, age 13)

Related Behaviors
While interviews focused largely on bullying, many of the participants mentioned that adolescents who are involved in bullying commonly take part in a range of other related behaviors. These individuals were described by a number of participants as "nakal," which translates to "naughty" or "delinquent." The most common behaviors attributed to adolescents who fell into this category were smoking, drinking, and fighting. In addition, participants mentioned driving motorcycles recklessly, skipping classes, skipping prayers, staying out late, dating inappropriately, and participating in "reog" (a traditional Javanese dance associated with mysticism).
Usually when they hang out to ride motorcycles together, the whole group is there. They may also smoke and drink together. When they get drunk, they usually don't think straight and usually fight each other. (Boy,age 13) Like bullying itself, there were gendered dimensions to nakal, with the behaviors described above most commonly associated with boys. While a number of participants mentioned girls who participated in these behaviors, this was generally seen as unusual, particularly as it pertained to physical aggression and substance use. Several adolescents suggested that girls who were nakal were more likely to participate in a somewhat different set of behaviors than their male counterparts, including things like forming gangs, dressing provocatively, swearing, and breaking rules. To the extent that these girls did engage in more "male-oriented" behaviors, this was depicted as largely a function of improper socializing with boys who were nakal. As the following adolescents articulated: It happens among girls as well, some are nakal, but not like that, not like the boys . . . For them, nakal is in the way they socialize with others . . . like being friends inappropriately, being friends with boys, nakal boys, like that. (Girl, age 14) One of my friends is my neighbor, he's been my friend since I was little. I would go out with him a lot, and I smoked, and I was drunk, too. I would go home at 11:30 pm, at 12:00 am . . . My [male] friends said to me, "You're a girl, why would you hang out with us, smoking and drinking like this?" (Girl, age 13)

Causes of Bullying
Regardless of gender, being nakal was considered to be a significant driver of bullying by many participants. Adolescents who fell into this category were characterized as disruptive and oppositional, frequently starting trouble with others for no apparent reason or as a way of seeking attention. For example, one girl profiled a boy that she labeled as nakal in the following way: One time, a girl didn't want to give up her seat to a boy. The boy forced her to do it. He didn't accept it, and spat on her . . . This boy does that a lot . . . and he also hits [her] . . . The boy is a nakal boy, and often gets into trouble with the school counselor . . . No one wants to be friends with him . . . [because] he's annoying like that . . . Like for example, he likes to hit others for no reason. And he writes on others' work, too . . . Yeah, he's a troublesome child and the school should do something about him. (Girl,age 14) Numerous participants also suggested that bullying was a way for "strong" adolescents to assert their power over "weak" adolescents. Not only were perpetrators described as being physically strong, but several participants also indicated that they were socially strong: adolescents with lots of friends who "rule the school," as one girl noted. Participants proposed that such adolescents perpetrated bullying as a way of demonstrating this power: "to gain notoriety so people will fear them" and "to let others know that they're strong." As a result, they often targeted those who "don't dare to fight back": adolescents who were commonly characterized as short, weak, nerdy, quiet, poor, or friendless.
There's this gang [of girls] in school . . . They say it's a gang only for people who have lots of friends, but if people don't have many friends you're not cool so you can't join . . . Most juniors and some seniors fear this gang. For example, if you make a little mistake, they'll confront you . . . They do whatever they want . . . they treat people how they want. (Girl,age 14) Beyond the abovementioned traits, participants noted a number of other characteristics that might cause an adolescent to become a victim of bullying. Physical appearance was a common theme, with those who were overweight or dark-skinned frequently targeted. Likewise, a number of participants mentioned physical and cognitive deficiencies: for example, having an overbite, speaking with a stutter, or having a cognitive disability. Those who were seen as gender nonconforming might also be bullied: this was particularly the case for boys who were perceived to be "girly," although several participants also mentioned girls being targeted for certain unsanctioned behaviors (e.g., smoking, drinking, being overly flirtatious).
He was mocked with "pretty" during a class, I forget which class, pretty [boy's name], pretty [boy's name], like that, and the boy looked like he wanted to fight back but he refrained . . . there was a point when he wrote something on the blackboard and he was really gentle with it, and the teacher said, "Oh my god, son, are you a girl or a boy? Why do you write so softly like that?" (Girl, age 13) There was this group of boys . . . There's one boy in my class that's really . . . like hollow, that's it. . .a little lacking. Well he's lacking in . . . if we rate him out of a hundred percent, he has less than that . . . The group of boys fought with him. They sent one person to fight that boy who was lacking. (Boy, age 13) According to several participants, another noteworthy driver of bullying was jealousy, with adolescents targeted for doing well in school or being involved in a romantic relationship. Finally, many participants suggested that adolescents often perpetrated bullying largely for their own entertainment and enjoyment. As one boy said, "They have nothing else to do, probably. They're just bored."

Consequences of Bullying
Throughout the interviews, participants frequently described both the emotional and social consequences of bullying. In terms of the emotional Fine et al.

NP4101
consequences, a consistent term used by adolescents was "sakit hati," which translates literally to "sick heart." Participants' descriptions of sakit hati suggest that this term encompasses a range of emotional responses, including hurt feelings, sadness, rumination, and depression. A number of adolescents suggested that sakit hati is a state that can last for an extended period of time, although it was clear that this was not always the case. Participants noted that sakit hati was the consequence of an experience that "becomes too much" and "enters the heart," and a few mentioned that in severe cases it could lead to suicidal behaviors.
[The victims] are sakit hati, and they want revenge but they can't get it because [the bullies] are more powerful. [They feel] sad, depressed, and scared, like they are burdened. (Girl, age 13) If it's sakit hati you can see it in their faces or their eyes . . . Their faces show some sort of fake expression . . . at school, if someone is joking around, they would just have a fake smile . . . the look in their eyes is as if they wanted to confide in others but they are buried inside themselves . . . they are afraid that if they tell others, later their words will spread to everyone. (Girl,age 13) Alongside sakit hati, participants mentioned a wide array of related emotions, including feeling sad, angry, ashamed, stressed, and depressed. Many also mentioned feelings of fear: for example, being afraid of the bullies themselves, afraid that the bullying would get worse if they tried to resist, and afraid that others would target them if they found out about the bullying. According to adolescents, many of these emotional consequences were related to the primary social consequence of being bullied: social isolation. Participants described how those who were sakit hati would intentionally distance themselves from their peers. Likewise, those who were afraid would choose to stay away from others to avoid instances of bullying. These decisions to stay away seemed intertangled with forced isolation due to ostracization, particularly in cases of relational bullying: I was sakit hati, and my friends started to avoid me. And then they told everyone not to be friends with people like that, like me . . . So, it was better for me to be quiet . . . I was sakit hati so I decided just to be alone. (Girl,age 13) Social isolation was also described as a consequence for the bullying perpetrator. Several participants detailed instances in which bullies became "hated for their behavior" and suggested that adolescents would alienate the bullies out of fear or dislike. Relatedly, a few participants mentioned the ways in which bullies would develop a bad reputation, being labeled as "bad" or "nakal" by their peers or teachers. One boy illustrated such consequences through his description of what happened to the bully in his group of friends: Even if we didn't make any mistakes, he sometimes beat us up . . . Yeah, he would boss us around . . . order us to do this and that . . . The disunity started when there was one person who didn't accept it any longer. And then, he asked another person to join him. Both of them agreed. Then they began to incite other friends to join them over time . . . [Now we] stay away from him . . . For example, if he needs something, we don't want to help him . . . No one starts conversations with him anymore. (Boy,age 13) Finally, participants mentioned a number of educational consequences for bullying perpetrators. Most commonly, adolescents would be called into the school counselor's office. In serious cases, the school counselor might then call the adolescent's parents into school for a meeting. A few participants also mentioned other punishments from the school, such as receiving a warning letter or getting suspended. Importantly, several adolescents suggested that while those caught bullying were often called in to speak with the school counselor, this intervention had little lasting impact: I've reported this to the school counselor before, but they just had a discussion with this boy, and it didn't really go through. I'm not satisfied with the response.
[The school counselor] only told the boy not to do it again, to do something else when he's joking around. His parents were called, and they didn't even come. (Girl,age 14) Coping Strategies Victims of bullying were described as coping with their experiences in a number of different ways. The most frequently cited coping strategy was sharing problems with close friends. Similarly, a number of adolescents mentioned seeking support from parents or teachers. By contrast, numerous participants said that adolescents are often reticent to share their experiences of bullying with others and would instead respond by "staying quiet." Participants cited a range of reasons that adolescents preferred to stay quiet. For instance, adolescents worried that if they told a teacher and the bully got in trouble, then the bullying would get worse. They worried that if they told a friend, others would end up knowing what had happened to them. They worried that their parents would be angry with them for being involved in bullying, even as a victim. Notably, both of these competing coping strategies-seeking support versus staying quiet-were much more commonly mentioned by girls compared to boys. I've told her several times to report him to the school counselor or try to fight back. She doesn't want to, because she's worried that if she responds like that it will be even worse for her. So she just succumbs to it, just stays quiets. (Girl,age 14) My parents would get angry, I think they would get angry, and later I would get blamed. So I just stayed quiet. (Girl,age 13) Another consistent theme throughout interviews was the feedback loop of interpersonal aggression, with those who were initially victims of bullying often becoming perpetrators either as a way of "venting their anger" or as a means of retaliation. Finally, many participants suggested that for boys, physical fighting was a common and acceptable way for coping with bullying victimization. Several adolescents related this to the idea of pride: if a boy "didn't accept" being mocked, particularly when the mockery was related to his family's reputation, then getting into a fight was an appropriate recourse. These fights were generally described as short-lived, with both parties moving on from the conflict afterwards.
Sometimes [a boy] can't accept it when his physical appearance is mocked, or when his parents are mocked. That's the main cause of two boys fighting . . . They'll punch and kick each other, and their friends will egg them on. But [the fighting] usually doesn't last long. Someone will try to separate it. (Boy, age 13)

Discussion
The current study explored Indonesian early adolescents' motivations, perceptions, and beliefs regarding bullying involvement. Consistent with previous qualitative studies (Guerin & Hennessy, 2002;Hellström et al., 2015;Vaillancourt et al., 2008), we found that Indonesian adolescents did not uniformly emphasize the traditional bullying criteria of intentionality, repetition, and power imbalance. Of these factors, power came up most frequently in interviews, with bullying commonly seen as a way for adolescents who were physically or socially strong to assert themselves over those they perceived as vulnerable. Repeated behaviors were also mentioned by a number of adolescents, particularly as an indication of when bullying crossed the line from innocuous to serious. Intentionality arose rarely, although a few participants again suggested that harmful intent could be a marker of serious bullying. While all three criteria emerged, however, their lack of consistent endorsement across the majority of adolescents is indicative that this widely used bullying definition may not be entirely appropriate within the Indonesian context. Further, while the overall inconsistences between youth-and researcher-generated descriptions of bullying are unsurprising, they add weight to recent arguments around the need for an updated definition with a greater basis in empirical evidence (Vivolo-Kantor et al., 2014;Volk et al., 2017).
Notably, in distinguishing between "normal" and "serious" bullying, a number of participants stressed that the harmful consequences for the victim outweighed the harmful intent of the perpetrator. For these adolescents, interpersonal aggression could be considered bullying to the extent that it caused significant emotional distress (e.g., sakit hati). Several other studies have documented similar findings around the importance of the victim's experiences in defining bullying (Guerin & Hennessy, 2002;Hellström et al., 2015). While a robust body of literature has established emotional distress as a consequence of bullying victimization (Moore et al., 2017), these studies suggest that such distress should be considered fundamental to the very nature of bullying. Despite this evidence, however, few definitions of bullying have explicitly included harmful consequences as a central criterion. A notable exception is Volk et al. (2014), who proposed a modified definition of bullying as "aggressive goal-directed behavior that harms another individual within the context of power imbalance." The authors argue that including harmful impact as part of this definition removes the conflation of repetition with harm, and instead acknowledges that a victim's perception of harm can be amplified by both the frequency and severity of bullying (Van Noorden et al., 2016;Ybarra et al., 2014). Findings from the current study lend support to this notion: adolescents indicated that the seriousness of bullying was intimately tied to the harm it caused, which could be related to bullying type or frequency.
Participants' accounts of the consequences of serious bullying underscore the importance of pinpointing a contextually appropriate definition for this phenomenon. Adolescents highlighted a range of short-and long-term emotional challenges resulting from bullying victimization, encapsulated by their descriptions of sakit hati. A few participants suggested that at its most severe, bullying involvement could lead to suicidal behaviors, an association that has been well-established by quantitative evidence (Heerde & Hemphill, 2019). It was also apparent that emotional distress is often intimately connected to social isolation, and that each of these negative consequences can amplify the other. This has particular salience when paired with the admission that many adolescents choose to "stay quiet" about their bullying experiences, as it illustrates the ways in which adolescents' responses to bullying may deepen their feelings of isolation. While studies have established associations between bullying and loneliness (Moore et al., 2017;Yusuf et al., 2019), future longitudinal investigations should tease apart the complex pathways linking bullying involvement, social isolation, and emotional distress among Indonesian youth.
Across qualitative interviews, a number of noteworthy gender differences emerged related to adolescents' involvement in bullying and other related risk behaviors. While both boys and girls participated in verbal bullying, adolescents suggested that boys were more likely to be involved in physical bullying and girls were more likely to be involved in relational bullying. This aligns with widespread views of gendered aggression, in which physical aggression is assumed to be more masculine and relational aggression is assumed to be more feminine (Crick et al., 2007;Juvonen & Graham, 2014); while studies have continually confirmed that boys are more likely than girls to engage in physical bullying, however, findings have been more equivocal regarding gender differences in relational bullying (Card et al., 2008;Casper & Card, 2017). Stark gender differences also emerged in boys' and girls' divergent expressions of other behavioral risks, with boys perceived as being predisposed to smoking, drinking, fighting, driving recklessly, and staying out late. Together, these findings speak to the ways in which underlying gender norms can drive vulnerability among adolescent boys, with normative values around masculinity supporting the adoption of aggression, delinquency, and substance use (Ragonese et al., 2019). Further, they suggest the need for gender sensitivity when building prevention programs targeting behavioral challenges among adolescents in this context.
Interviews also illuminated boys' and girls' unique strategies and vulnerabilities related to coping with bullying experiences. Girls often relied on social support from friends or family in order to cope with bullying victimization and its resultant emotional distress; this reflects prior evidence which has established that girls have a greater tendency to seek support compared to boys (Hellström & Beckman, 2020;Hunter et al., 2004). At the same time, many girls expressed feeling afraid or uncomfortable with sharing their bullying experiences, choosing instead to stay silent. By contrast, boys commonly used physical outlets in order to cope with bullying, starting fistfights as a means of resolving interpersonal conflict. Participants' descriptions of these fights suggest their widespread acceptability: a socially sanctioned method for boys to vent their anger and defend their pride. While this type of coping may present a short-term solution for distressed boys, however, it is concerning due to the ways in which it normalizes violence as an acceptable problem-solving strategy, potentially foreshadowing the use of aggression within intimate relationships. Indeed, an emerging body of research has established links between bullying involvement and subsequent dating violence perpetration (Zych et al., 2019).
Finally, qualitative interviews yielded important contextual insights into the interpretation of the four previously established latent classes. In particular, adolescents' differentiation between normal and serious bullying provides a useful lens through which to understand the distinction between the Maladjusted and Behavioral Problems subgroups: the former in which bullying involvement is accompanied by a range of emotional problems, and the latter in which it stands on its own. Given the abovementioned qualification that serious bullying can be defined by its resultant emotional distress, it follows that bullying involvement may be more problematic among youth within the Maladjusted class compared to those within the Behavioral Problems class. A number of researchers have noted the normative nature of bullying among school-aged youth (Salmivalli, 2010;Smith & Brain, 2000). This was certainly reflected by participants in the current study, many of whom described bullying as a commonplace occurrence within school settings, a typical part of interpersonal dynamics, and a way for youth to entertain themselves and others. This aspect of bullying can make it particularly insidious from an intervention standpoint due to the challenges of shifting behaviors that are reinforced by social norms (Juvonen & Graham, 2014). Our mixed methods results suggest a possible entry point into this issue by allowing for the identification of those for whom these widespread behaviors are most consequential.
Together, these findings can be used to inform bullying prevention programs targeting Indonesian youth. The normative nature of bullying in this context speaks to the need for whole-school approaches which attempt to shape the broader social climate in schools through a combination of components targeting entire schools, classrooms, teachers, and individuals (Vreeman & Carroll, 2007). These approaches have shown promise in high-income contexts, with several systematic reviews suggesting that they are more effective in reducing bullying than those solely focused on individual victims or perpetrators (Cantone et al., 2015;Silva et al., 2017;Vreeman & Carroll, 2007). Notably, a recent study piloted a whole-school intervention aimed at shifting social norms around bullying in Indonesian secondary schoolsincluding four schools in Semarang-and found this approach to be both feasible and acceptable (Bowes et al., 2019). Despite the intervention's promise, however, findings were mixed in terms of its impact on bullying prevalence, although it is important to note that the study was underpowered to detect statistical changes. The authors concluded that while social norms around bullying were a potentially malleable intervention target, future contextual adaptations were warranted.
Our qualitative results serve as a roadmap for the types of adaptations that may be necessary in order to more effectively reach Indonesian youth. First, they imply the need to acknowledge adolescents' own definitions of bullying, especially given inconsistencies with traditional conceptualizations of bullying. For instance, given the centrality of the victim's experience of emotional distress to adolescents' understanding of problematic bullying, anti-bullying messaging could focus on its harmful consequences in order to emphasize the ways in which this normative behavior can become serious. Second, our results suggest that interventions may not be equally applicable to both boys and girls and should therefore consider gender differences in bullying involvement as well as the role that underlying gender norms play in driving these differences. Finally, the fact that many adolescents were unwilling to disclose significant bullying experiences underscores the importance of building better psychosocial support systems within school settings. While school counselors are intended to provide this support, they were largely described by participants in terms of their disciplinary roles-and many viewed this discipline as ineffective. As such, there is a clear need to equip school personnel with the skills to better address emotional distress among vulnerable students.
A number of study limitations must be discussed. First, as mentioned previously, our original analytic plan was to make thematic comparisons across adolescents from the four latent classes. This was ultimately infeasible due to variations in data richness: while some participants were excellent informants, some had immense difficulty articulating their thoughts and experiences, which is a challenge inherent in conducting qualitative research with younger adolescents (Huang et al., 2016). As such, we have not indicated participants' latent class assignments for the selected quotations, nor do we attempt to attach differential meanings to qualitative data drawn from participants from separate classes. Instead, we have largely relied on class assignments as a means of ensuring representation from adolescents with diverse emotional and behavioral experiences. Second, findings from qualitative interviews conducted in Semarang may not be transferable to adolescents across Indonesia, particularly given the country's immense ethnic, cultural, and linguistical diversity. Future qualitative studies should replicate these efforts with Indonesian youth drawn from a range of different environments in order to help inform country-wide bullying prevention efforts. That said, the fact that our findings conform to those from other communities suggests that a number of the experiences young people report may be more common than divergent. Third, while interviews were conducted in Bahasa Indonesia, the majority of data analysis was undertaken in English. We have attempted to address this issue through a bilingual process of open coding as well as ongoing collaboration with Indonesian partners; it is still likely, however, that relevant information was lost in translation.
Despite these limitations, our results demonstrate the utility of exploring adolescent bullying involvement using a mixed methods framework. Building on prior quantitative research-which uncovered four distinct patterns of emotional and behavioral problems among Indonesian youth-qualitative findings emphasized the myriad ways in which bullying involvement ties into other psychosocial challenges. Overall, results highlighted the need to consider youth's own perceptions of bullying in order to develop effective prevention strategies in Indonesian schools.