Explaining Suicidal Behavior Among Adolescents in the Urban Community Context: Insights From a School‐Based Study in Depok City, Indonesia

ABSTRACT

Suicidal behavior among adolescents is a growing global mental health concern. In addition to underreporting in the Global South or low‐middle‐income countries, studies on suicidal behavior among adolescents are still limited. Moreover, suicidal behavior can be triggered by various factors related to rapid urbanization, but there is a lack of studies exploring the determinants of suicidal behavior among adolescents in specific urban community context in the Global South. The present study aimed to describe the suicidal behavior among adolescents in an urban community in Indonesia, as well as explain the factors contributing to such a public health problem. This cross‐sectional study is based on the data from 446 respondents who participated in an online School Climate Survey at two public junior high schools in Depok City, Indonesia. Measured using a composite suicidal behavior index, around 7% participants (n = 29, the majority of whom were female) reported ever having suicidal thoughts, plans, or attempts. Variables related to student‐parent relationships, school and social support, and exposure to antisocial behavior were assessed for their effect on suicidal behavior. A logistic regression analysis revealed that being an adolescent female, the amount of time spent with family, exposure to physical assault, and truancy predicted suicidal behavior in adolescents. The study adds to the existing limited body of knowledge on suicidal behavior among adolescents in the urban community context of countries in the Global South and provides suggestions for developing tailor‐made interventions to specific adolescent characteristics.

1. Introduction

There is an alarming increase in suicide rates among young people around the world in the last decade. Suicidal behavior is commonly conceptualized as two distinct but related components: suicidal ideation and suicide attempts. Suicidal ideation can be defined as thoughts or wishes to end one's life, while suicide attempts are defined as self‐injurious behaviors intended to result in death (Lewinsohn et al. 1996). Importantly, a suicide attempt is any self‐directed injurious behavior with the intent to cause injury, with or without resulting injury, whereas suicidal ideation is the thought, consideration, or planning to die by suicide (Crosby et al. 2011). Suicide ideators (Beck et al. 1972) are individuals who currently report suicidal ideation, including intentions and plans to die by suicide, but who have not recently made an overt suicide attempt.

Empirical data illustrate the global scope of this concern. Based on the US National Youth Risk Behavior Survey, suicidal ideation among youth shifted from a declining to an increasing trend beginning in 2009; by 2019, prevalence had decreased to 15% from 19% in 1991, while non‐fatal suicide attempts slightly increased from 7% to 9% (Xiao et al. 2021). Similar trends have been documented in Europe, where 69% of adolescents who considered suicide eventually attempted it (Eckhoff et al. 2020), and in the Caribbean, where prevalence rates of 26% for ideation, 26% for plans, and 19% for attempts were reported—with bullying victimization, cyberbullying, loneliness, truancy, and worry identified as key risk factors (Sarfo et al. 2022). In Asia, 26% of Japanese adolescents aged 13–18 reported suicidal ideation and 5% reported attempts (Nagamitsu et al. 2020), while in China, 24% of junior high schoolers reported ideation and 4% reported attempts (Peng et al. 2019)—collectively suggesting approximately 1 in 12 adolescents has attempted suicide, lower than rates observed in Western countries. Across Africa, a meta‐analysis estimated adolescent suicide attempt prevalence at 10% (Babajani et al. 2024).

A relatively new study in the Southeast Asian (i.e., Thailand, Indonesia, Timor Leste) context reported suicidal ideation and attempts among adolescents aged 12–15 years varied from 5% to 13% and 3% to 11%, respectively (Li et al. 2021). In Indonesia, 4% of Indonesians aged 13–15 reported experiencing suicidal ideation according to the 2007–2013 Global School‐based Student Health Survey. This figure placed Indonesia sixth among the seven countries, slightly above Myanmar, which recorded the lowest prevalence at 1% (Peltzer et al. 2017). Recent studies revealed a similar pattern, showing that 5% of Indonesian youth surveyed seriously considered attempting suicide, and 3% attempted suicide in the last 12 months (Putra et al. 2021).

A substantial body of research indicates that suicidal ideation increases with age throughout adolescence and tends to stabilize as young people enter adulthood (Lewinsohn et al. 1996; Geoffroy et al. 2016). Consistent with this pattern, the World Health Organization (2021) identifies suicide as one of the leading causes of death among individuals aged 15–29 years. Importantly, the WHO report highlighted the problem of underreporting in low‐middle‐income countries, which may be attributed to the lack of suicide prevention efforts that these countries (i.e., India, Bangladesh, Indonesia, and Sri Lanka) have implemented (Arafat et al. 2020). The underreporting of suicide may also be influenced by the Indonesian's deeply rooted religious values and the stigma attached to suicide, which discourages individuals from openly disclosing their experience with suicide in self‐report surveys (Putra et al. 2021). In Indonesia, while suicide attempts are not formally criminalized, religious and customary norms widely regard suicidal behavior as a violation of the sanctity of life and prevailing moral order (Widyawati and Qurtuby 2025). This framing tends to cast individuals experiencing suicidal ideation as morally or religiously deficient, compounding shame and social stigma while discouraging help‐seeking and underreporting of suicidal behavior.

Previous studies have explored the risks and determinants of adolescent suicide in general. These include peer victimization (Geoffroy et al. 2016), residing in a low‐middle‐income country, lack of friends, loneliness, bullying victimization, having been in a physical fight in the past 12 months, lack of parental or guardian support, and alcohol and tobacco use (Peltzer et al. 2017). Additional factors include lack of physical activity, poor sleep, the school environment (Pfledderer et al. 2019), teen dating violence (Baiden et al. 2021), truancy, physical fighting, dietary deficiencies (rarely eating fruit), sedentary behavior, and being a female (Marthoenis and Yasir Arafat 2022). Given the limited research and the lack of efforts to prevent suicide in Indonesia, this study adds to the literature on this important topic.

The present study utilized data from school climate surveys in two schools in Depok City in Indonesia, a country ranked as the fourth most populated country and the fifth largest urban population in the world (approximately 162.56 million in 2023). Located within the metropolitan region surrounding Jakarta, the northern part of Depok City shares a direct border with Jakarta, the capital of Indonesia, and it is affected by the urban sprawl, leading to a high population density. Badan Pusat Statistik/BPS‐Statistics Indonesia (2024) recorded a population density in Depok City of 9644 people/km², with 2% of its population below the poverty threshold. Depok City is experiencing rapid urbanization challenges with an influx of people. The bustling city has several poor urban settlements which are characterized by poverty, poor access to schooling, lack of opportunities for youth, drug and alcohol abuse, domestic violence, organized violence and criminality. Similar demographic and structural conditions have been increasingly observed in emerging urban areas throughout Indonesia (Khairunnisa et al. 2024; Niman et al. 2025; Sujarwoto et al. 2026) and Southeast Asia (Cimene et al. 2022). As such, Depok City represents a valuable case for understanding adolescent suicide risk in rapidly urbanizing contexts, generating insights that may be relevant not only to other Indonesian cities but also to comparable urban settings across the region.

1.1. The Suicidal Behavior Theory

The Interpersonal Theory of Suicide (IPTS; Joiner, 2005) has become one of the most influential yet contested theories of suicidal behavior (Hjelmeland and Loa Knizek 2020; Smith et al. 2022). Its influence derives largely from its reduction of suicidal behavior to three key interpersonal constructs: acquired capability for suicide, thwarted belongingness, and perceived burdensomeness. The theory is one of the first to utilize an ideation‐to‐action framework (i.e., distinguishing between those who only have thoughts and those who have thoughts followed by an attempt) highlighting a transition from suicidal thoughts to suicidal actions that may occur when the three constructs present simultaneously, and has been discussed in the context of military veterans (Compton et al. 2021), college students (Zhang et al. 2013), and transgender and gender nonconforming individuals (Grossman et al. 2016), among others. Critics, however, argue that this framework oversimplifies the complexity of suicide by placing insufficient emphasis on the broader social, cultural, and contextual factors that shape suicidal experiences.

The Three‐Step Theory of Suicide (3ST), proposed by Klonsky et al. (2016; 2021), offers an empirically grounded framework for understanding the development of suicidal ideation and behavior. According to the theory, the presence and intensity of the suicidal ideation are driven by the interaction of pain, hopelessness, and (loss of) connectedness. The third step is the progression from ideation to action (suicide attempt). Importantly, and in contrast to IPTS, researchers argue that the theory should not be viewed without contextual understanding of factors that shape suicidal experiences. Recent literature in the nursing field (Anderson and Happ 2021) examining this theory also highlights the necessity of considering contextual factors in a specific population.

Growing evidence suggests that suicide is a multifaceted phenomenon shaped by interacting social, psychological, cultural, biological, and environmental factors (Wang et al. 2020; Prades‐Caballero et al. 2025). Adopting Bronfenbrenner's Ecological Systems Theory (1977; 1979), multilevel perspective evaluates suicide risk by mapping layers of the individual's environment to understand how they mutually interact (Yang et al. 2025). Beyond interpersonal and ideation‐to‐action theories, adolescent suicidal behavior is therefore best understood through a socio‐ecological lens, which places adolescents within interconnected environments that generate the risk and protective factors associated with suicidal behavior.

1.2. Contextual Factors Influencing Adolescent Suicidal Behavior

Previous literature has examined how school and family environments are related to adolescent suicidal behavior. For example, using cross‐sectional data from the 2013 Texas Youth Risk Behavior Survey, researchers found that the more adolescents spent their time at home with their parents, the less likely they were to have considered suicide and to have attempted it (Olcoń et al. 2017). Similarly, more parental involvement in supporting their 13–15‐year‐old children helped to reduce their children's suicidal thoughts in Southeast Asian countries (Peltzer et al. 2017). In a recent study, South Korean adolescents living without parents had significantly higher suicidal ideation and reported more attempts, and this also affected their overall likelihood of suicidal ideation (Jeong et al. 2020). Similarly, in Indonesia and neighboring countries, the presence of parental supervision has been found to play a protective role in actually reducing the risk of suicide (Marthoenis and Yasir Arafat 2022).

Feeling unsafe at school could further increase their suicidal ideation and the likelihood of attempting suicide (Nickerson and Slater 2009). Feeling safe at school is one indicator of school connectedness (Marraccini and Brier 2017), along with other school climate indicators such as feeling cared for/respected by teachers and perceived availability to interact with teachers, among others. Marraccini and Brier (2017) also found that higher levels of school connected were not only seen to reduce the likelihood of suicidal ideation among high‐risk (e.g., sexually abused, bullied) populations, but among the general student population as well. The relationship between the school environment and suicidal behavior has also been observed across diverse geographic regions (Marraccini and Brier 2017; Madjar et al. 2018; Han et al. 2018; Zheng et al. 2023). In the Middle East, teacher support—defined as facilitating an effective learning atmosphere in the classroom—was significantly related to a lower rate of suicidal ideation (Madjar et al. 2018). In the same study, positive support from parents and peers had a similar impact on the reduced likelihood of suicidal ideation among the students. Evidence from a study on Chinese adolescents in the school setting showed a significant association between the student‐teacher relationship dissatisfaction and an increased likelihood of having suicidal thoughts (Zheng et al. 2023). Additionally, one study demonstrated that positive student‐teacher relationships served as a protective factor against both suicidal thoughts and attempts, even in the presence of bullying (Han et al. 2018). These findings underscore the critical role of supportive interpersonal relationships within the school setting in mitigating suicide risk among adolescents.

Suicidal behavior is also strongly associated with violence and bullying. A study by Peltzer et al. (2017) highlighted that exposure to physical fights and being a victim of bullying are significantly related to suicidal ideation among adolescents in Southeast Asian countries. Similarly, Putra et al. (2021) found that the experience of being physically attacked and bullied are strong predictors of suicidal behavior among females and males. A similar trend was also observed in the United States, where being threatened and bullied at school increased the likelihood of attempting suicide (Olcoń et al. 2017). The relationship between victimization and suicidal behavior is striking. Peer victimization has been identified as a significant predictor of suicidal ideation at age 13 and subsequent suicide attempts within the following 2 years, even after controlling for factors, such as mental health problems, socioeconomic status, intelligence. (Geoffroy et al. 2016).

While it is widely known that female adolescents are more susceptible to suicidal behavior (Putra et al. 2021; Jeong et al. 2020; Geoffroy et al. 2016; Nock et al. 2008; Peltzer et al. 2017; Marthoenis and Yasir Arafat 2022), teen dating violence is another contextual factor associated with suicidal behavior. In a multivariate model controlling for persistent sadness, sexual orientation, and significant risk behaviors, recent teen dating violence increased the odds of suicide attempts by 1.61 for female adolescents (Olshen et al. 2007); the effect was not observed for male adolescents. Recent literature further supports these findings, indicating that adolescents who experienced physical dating violence were 1.92 times more likely to have thought about suicide, 1.67 times more likely to have made a plan, and 2.42 times more likely to have attempted suicide (Baiden et al. 2021).

Importantly, substance use is also associated with suicidal behavior. In both univariate and multivariate analyses (Chan et al. 2014), substance use (e.g., smoking, drinking alcohol) has been identified as a strong predictor of suicidal behavior. This finding is consistent with prior research conducted in Australia (Beautrais 2000) and across Southeast Asia (Peltzer et al. 2017), where substance use and antisocial behavior are associated with elevated suicide risk among adolescents. The effect was also observed among students who skipped class (Marthoenis and Yasir Arafat 2022) in Bangladesh and Indonesia, along with other factors mentioned above (i.e., those who experienced being a victim of aggression, bullying, mental health challenges, and substance use—i.e., cigarettes, alcohol). According to a recent systematic review, truancy (i.e., being absent from school) is also one of the markers of social exclusion that is associated with increased self‐harm and suicidal ideation (Epstein et al. 2020), although the mechanism underlying this association remains unclear.

While there has been enormous discussion of suicidal behavior from the lens of intrapersonal issues (i.e., anxiety, depression), the aim of the present study focused on contextual influences and explored the relationships between these factors and suicidal behavior (ideation and attempt). The contextual factors investigated included: student‐parent relationship (i.e., bonding and time with family, parental support of academic achievement, parental control and monitoring), positive school climate (i.e., perception of school safety, teacher support, perceived availability of student‐teacher interaction, positive relationships with peers), and problem behaviors (i.e., exposure to physical fights, bullying victimization, experience of being physically attacked, being threatened or injured, teen dating violence, substance use, truancy). Specifically, the study sought to identify which factors were associated with and increased the odds of suicidal behavior.

2. Method

2.1. Participants

The study involved students from two junior high schools in Depok City, referred to as School Y and School Z—to maintain confidentiality due to the sensitive nature of the data. These schools were recommended by the local office of education and signed a memorandum of understanding with Faculty of Psychology, Universitas Indonesia to participate in a large project entitled “Addressing the wellbeing and security needs of urban children and adolescents in Indonesia in the digital era: adapting the Safe Communities Safe Schools (SCSS) Model in violence‐prone urban areas in Indonesia” funded by Fondation Botnar, Switzerland, from 2019 to 2026. This publication was developed as part of the project implemented by Social and Crisis Intervention Lab, Universitas Indonesia in collaboration with Danish Institute Against Torture, and Center for the Study and Prevention of Violence, University of Colorado Boulder, and Universitas Syiah Kuala.

Sample size was determined by using G*Power 3.1.9.7 (Faul et al. 2009) for an a priori power analysis for a logistic regression. The required sample size was calculated to detect a moderate effect (odds ratio [OR] = 2.0) with an event probability of 3% (Putra et al. 2021), power of 0.80, and alpha level of 0.05. The analysis indicated that a minimum of 567 participants was required. To ensure a representative sample across grade levels and gender, a stratified random sampling method was employed. However, only a total of 552 students [298 out of 810 eligible students (37%) for School Y and 254 out of 720 eligible students (35%) for School Z] consented and participated in the online survey, which was conducted in a classical classroom setting. To ensure that there were no careless responses due to the length of the 413‐item self‐report questionnaire, the study filtered responses based on the duration of the survey at the end of the pilot phase of the survey adaptation (M _duration = 53.29 min, SD = 18.43). Careless response is common in an online survey (Johnson 2005) and when completing a long questionnaire, especially toward the end of the survey (Berry et al. 1992). Thus, we selected responses completed within 34.86–71.72 min, and eventually dropped 106 participants. The final number of responses retained for further analysis was 446.

2.2. Procedures

The data collection was conducted on the full sample in December 2022. In each school, participants were distributed across 10 classrooms with a researcher to ensure that participants completed the online survey correctly. The researcher instructed students 1 week prior to data collection to bring their smartphones in order to access the survey online through a licensed survey platform called Alchemer (www.alchemer.com). The surveys were anonymous, and no identifying information was collected from students. Upon completion, students showed the submission page (indicating that they had completed the survey) to the researcher outside the classroom. Each student received a lunch box (valued at US$3) as an incentive for participation.

2.3. Measures

The present study used data from a cross‐sectional study, using an adapted version of the SCSS School Climate Survey‐Student Middle School version (n _items = 413) (Dymnicki et al. 2021). It was further translated and back‐translated, and pilot‐tested with 50 students in May 2022 for the specific urban context in Depok City. The pilot‐testing of the adapted version of School Climate Surveys showed an average Cronbach's alpha of 0.69. For certain sensitive questions (i.e., “Have you ever been in trouble with the law?”, “In the past 12 months, have you ever seriously considered attempting suicide?”, etc.), we provided a “refuse to answer” option so that if participants did not feel comfortable answering openly, they could opt out of responding by using the “refuse to answer” option instead of skipping the question, resulting in ambiguous missing data. These changes were also made following the recommendation of the Ethics Committee of Faculty of Psychology, Universitas Indonesia, which granted ethical approval for the study (No. 165/FPsi.Komite Etik/PDP.04.00/2022). The final version consisted of 292 items under 57 subscales (22 subscales using a Likert‐type response set, with an average Cronbach's alpha of 0.82) measuring school climate, risk and protective factors for school violence, and violent and other problem behaviors. The study used a self‐report survey, which included a required “refuse to answer” option (as suggested by the ethics committee); while this may have reduced the at‐risk rate by allowing individuals to withhold sensitive answers, the survey's anonymous format and the inclusion of such a response can increase participation (Ong and Weiss 2000).

Because this study involved junior high school students, who are in early adolescence and under the age of 17, both legal custodians of the students provided written consent for their child to participate in the SCSS program and research project. Students also provided assent, and they could opt out of the study themselves. Therefore, this study only included data from participants who gave their assent along with their parents’ consent.

2.4. Student Suicidal Behavior

Suicidal behavior was measured using a similar approach to Baiden et al. (2021), who conceptualized it through three items with categorical response options (see Table 1).

Table 1.

Questions measuring suicide behavior.

Items	Response set
1. During the past 12 months, did you ever seriously consider attempting suicide?	No, yes, refuse to answer
2. During the past 12 months, did you make a plan about how you would attempt suicide?	No, yes, refuse to answer
3. During the past 12 months, how many times did you actually attempt suicide?	0, 1 time, 2–3 times, 4–5 times, 6 or more times, refuse to answer

Items #1 and #2 represented suicidal ideation. Item #3 reflected a suicide attempt. The questions were presented logically—a skip pattern was created so that items #2 and #3 appeared if individuals responded “yes” to item #1. Yes and no responses in items #1 and #2 were re‐coded as 1 and 0, respectively. Similarly, item #3 was re‐coded as a binary response, where any response other than 0 was re‐coded as 1. The suicide risk score was then calculated with a range of 0–3, where 0 indicates no‐risk individuals and 3 indicates high‐risk individuals—that is, they responded “yes” in item #3 or all 3 items. Twenty‐four individuals refused to answer this subscale. Table 2 shows the percentages at risk for suicidal behavior.

Table 2.

The risk of suicidal behavior.

	All participants		Female		Male
Suicidal behavior	n	%	n	%	n	%
No‐risk	378	90	195	85	183	95
Low‐risk	10	2	5	2	5	3
Medium‐risk	5	1	5	2	—	—
High‐risk	29	7	24	11	5	3

2.5. Student‐Parent Relationship

The SCSS School Climate Survey‐Student version included three subscales to measure the student‐parent relationship. Well‐spent Time with Family (α = 0.75) assessed youths’ communication with parents and shared family activities, with items such as “I can tell my parents the way I feel about things,” “I like to do things with my family,” and “I usually have dinner with my family.”

Parental Support of Academic Achievement (α = 0.79) assessed the extent to which parents are involved in and encourage their child's education, with items such as “My parents make sure I do my homework,” “My parents want me to get good grades,” and “My parents talk to me about doing well in school.”

Parental Control and Monitoring (α = 0.74) assessed the degree to which parents monitor and are aware of their child's activities and social connections, with items such as “My family has rules about where I can go and what I can do,” “When I am not home, one of my parents knows who I am with,” and “My parents know who my friends are.” All subscales were scored using 4‐point Likert‐type responses (1 for “strongly disagree” and 4 for “strongly agree”).

2.6. Positive School Climate

Several subscales of Positive School Climate were used in the analysis. Teacher Support (α = 0.65) assessed students’ perceptions of how much teachers and school staff care about and support them, with items such as “Teachers at your school treat students with respect,” “Teachers at your school try to help students when they are having problems,” and “The principal and the rest of the school staff try to make your school a place students like to be.” These items were scored using 4‐point Likert‐type responses where responses ranged from 1 (“almost none of them do”) to 4 (“almost all of them do”).

Positive Peer Support (α = 0.74) assessed the quality of students’ social connections and sense of belonging with peers, with items such as “I have a friend my age who cares about me,” “My friends are interested in what I think and how I feel,” and “When I am at school, I feel lonely” (reverse coded). Similarly, the Student‐Teacher Perceived Interaction Availability factor was assessed using five items (α = 0.66): “There are lots of chances for students in my school to talk with a teacher one‐on‐one,” “I think the staff at this school do a good job talking with us about mental health problems,” “I feel like I could talk to my teacher or someone at my school if I was having a mental health problem,” “There is a staff member I can talk to here if I was worried about someone else's mental health problem,” “I have seen teachers/staff help kids with mental health problems at school.” All items were recorded in 4‐point Likert‐type responses (1 for “strongly disagree” and 4 for “strongly agree”).

Lastly, a single item, “During the past 1 month, on how many days did you not go to school because you felt you would be unsafe at school or on your way to or from school?” was used to assess students’ perceptions about their safety at school. The response was scored as a frequency (where 0 = “0 days,” 1 = “1 day,” 2 = “2 or 3 days,” 3 = “4 or 5 days,” and 4 = “6 or more days”) and then re‐coded as a binary variable where 0 represented “0 days” and 1 represented “any mentioned days.”

2.7. Problem Behaviors

Multiple measures of problem behavior were used, which measured students’ exposure to or participation in behaviors that could result in significant negative personal, social, academic, or legal consequences for the youth. Aggression was assessed using two items related to physical aggression at school. Exposure to physical fights was measured with the item”During the past 12 months, how many times were you in a physical fight on school property?” Responses were recorded by frequency, with 0 times coded as 0, and 1–12 times distributed across 7 response options. For analysis, responses were re‐coded into a binary variable: 0 (0 times) and 1 (any mentioned incidents). Similarly, being threatened or injured was measured with the item”During the past 12 months, how many times has someone threatened or injured you with a sharp weapon such as a knife, club, motorbike gear, iron ruler, sickle, or katana (Samurai's sword) on school property?” This item was recorded and re‐coded using the same approach.

Victimization from aggression and teen dating violence was assessed using several binary (yes/no) items. Experience of being physically attacked was measured with the item “Have you been hit by another student trying to hurt you?” Teen dating violence was assessed with two items: “During the past 12 months, did your boyfriend or girlfriend ever hit, slap, or physically hurt you on purpose?” and “During the past 12 months, have you been physically forced to have sexual intercourse when you did not want to?” Truancy was also measured using a yes/no item “During the past month, I ‘cut’ or skipped classes without an excuse.”

To assess Substance Use, students were asked about their past 30 days’ use of smoking, vaping, alcohol drinking, marijuana usage, illegal drugs (i.e., ecstasy, heroin), and illegal use of prescription drugs (i.e., painkillers such as Oxycontin, Codeine, Percocet). Responses were recorded as frequencies which were then re‐coded into yes/no responses (0 and 1).

Finally, the survey's Bullying Victimization scale was used to assess students’ experiences being bullied. The scale consisted of 12 items asking about specific types of bullying behavior (i.e., being called mean names, being left out of things on purpose, cyberbullying) experienced in the past couple of months. Responses were recorded as frequencies (not in the past couple of months, only once or twice, 2 or 3 times a month, about once a week, several times a week, never in my life) and re‐coded into yes/no responses (0 and 1).

2.8. Analytic Plan

All statistical analyses were performed using R (R Core Team 2023). After descriptive statistics were conducted, a Pearson correlation was performed using the cor.test() function (with the addition of ‘method = pearson’ in the syntax) to determine any association between independent variables with Likert‐type responses and suicidal behavior. For correlations involving categorical responses, a point‐bi‐serial correlation was used using the same cor.test() function (R Core Team 2023). Logistic regression analysis was used to predict suicidal behavior using the glm() function (R Core Team 2023), and confint() to obtain confidence interval and exp(coef()) to extract ORs. We also used anova(x, y, test = “Chisq”) function to compare model fit between regression models.

3. Results

3.1. Descriptive Statistics

The final number of youth surveys retained for further analysis was 446 (M _age = 13.56, SD = 0.95), with 55% females and 45% males from 1st year (37%), 2nd year (35%), and 3rd year (29%) of the selected junior high schools. In general, students reported that they spent a moderate amount of time with their parents (M = 2.90, SD = 0.60) and that the parents showed considerable support of their academics (M = 3.18, SD = 0.45). Correlation analysis indicated that more well‐spent time with family, r(444) = −0.23, p < 0.001, and higher parent support of academics, r(444) = −0.17, p < 0.001, had a significant negative association with suicidal behavior. However, although students reported sufficient parental supervision (M = 3.05, SD = 0.48), this variable was not statistically associated with suicidal behavior, r(444) = −0.09, p = 0.07. Students also described notable support from their teachers (M = 3.10, SD = 0.62) and peer groups (M = 3.06, SD = 0.45), as well as a fair level of ease in interacting with their teachers (M = 2.90, SD = 0.42). The high level of teacher support, r(444) = −0.14, p = 0.002, peer support, r(444) = −0.15, p = 0.001, and teacher approachability, r(444) = −0.11, p = 0.02, were also associated with a lower risk of suicidal behavior. While a notable proportion of students reported feeling unsafe at school (20%), this factor was not significantly associated with suicidal tendencies.

In terms of victimization among students, respondents reported low incidences of unwanted sexual intercourse in a relationship (1%), romantic partner violence (2%), receiving threats (3%), school avoidance (9%), and witnessing or participating in physical fights (12%), but a considerable proportion reported being assaulted (23%). All of these variables were positively associated with a higher risk of self‐destructive behavior: unwanted sexual intercourse, r(444) = 0.22, p < 0.001; romantic physical abuse, r(444) = 0.13, p = 0.01; receiving threats, r(444) = 0.15, p = 0.001; school avoidance, r(444) = 0.23, p < 0.001; exposure to a physical fight, r(444) = 0.12, p = 0.01; and physical assault, r(444) = 0.20, p < 0.001. Additionally, there was a notable proportion of students reporting being a victim of bullying (20%), and low prevalence of prescription drugs misuse (3%), as well as alcohol (1%) and illegal drugs (less than 1%) consumption. These prevalences are also associated with suicidal threats: bullying victimization, r(444) = 0.25, p < 0.001; illegal drug uses, r(444) = 0.17, p < 0.001; drugs misuses, r(444) = 0.10, p = 0.03; alcohol consumption, r(444) = 0.19, p < 0.001. Finally, 8% of students reported vaping and 9% reported smoking; however, these behaviors were not associated with suicidal behavior.

3.2. Predicting Student's Suicidal Behavior

A binary logistic regression was computed to determine which variables strongly affected suicidal behavior. For this analysis, the suicidal behavior variable was re‐coded into two binary responses: 0 = no risk and 1 = risk. Additionally, due to very low incidence, alcohol consumption and illegal drugs use were not included in the model to avoid the potential for second‐order bias (Hahn et al. 2024). The predictors entered significantly improved model fit compared to a null model, χ ²(14) = 75.49 p < 0.001, and accounted for 33% of the variance in suicidal behavior. Table 3 presents the logistic regression coefficients (B), standard errors (SE), ORs, and p‐values, along with 95% confidence intervals, for each predictor.

Table 3.

Logistic regression of predictive variables to suicidal behavior.

Variables	B	SE	OR	p	95% CI
(Intercept)	0.24	1.74	1.27	0.90	[0.04, 36.56]
Female	1.52	0.48	4.56	0.001**	[1.90, 12.45]
Well‐spent time with family	−0.88	0.34	0.41	0.01*	[0.21, 0.81]
Parental support of academic achievement	−0.34	0.49	0.72	0.50	[0.27, 1.87]
Teacher support	−0.08	0.33	0.92	0.80	[0.49, 1.75]
Positive peer support	−0.38	0.45	0.68	0.39	[0.28, 1.67]
Student‐teacher perceived interaction availability	0.16	0.53	1.18	0.76	[0.42, 3.36]
Exposure to a physical fight	0.13	0.55	1.14	0.81	[0.36, 3.24]
Being threatened	1.34	0.85	3.83	0.12	[0.65, 19.28]
Experience of being physically attacked	0.95	0.42	2.58	0.02*	[1.12, 5.83]
Dating violence—physical abuse	0.51	0.98	1.67	0.60	[0.23, 10.96]
Dating violence— forced to have sexual intercourse	1.14	1.26	3.14	0.36	[0.31, 50.42]
Truancy	1.34	0.50	3.80	0.01**	[1.40, 9.95]
Bullying victimization	0.65	0.44	1.91	0.14	[0.79, 4.47]
Prescription drug misuse	1.08	0.95	2.96	0.25	[0.34, 16.21]

^* p < 0.05

^** p < 0.01.

As shown in Table 3, female students were 4.56 times more likely to be associated with suicidal behavior, compared to male students, OR = 4.56, 95% CI = 1.90–12.45, p = 0.001, indicating the strongest association with the dependent variable. Similarly, spending more time with parents significantly reduced suicide risk to nearly half OR = 0.41, 95% CI = 0.21–0.81, p = 0.01. Importantly, physical assault, OR = 2.58, 95% CI = 1.12–5.83, p = 0.02, and skipping school, OR = 3.80, 95% CI = 1.40–9.95, p = 0.01, had a strong association with suicidal behavior. The model shows that being female adolescents and skipping school have markedly higher odds of suicidal behavior.

4. Discussion

This study examined multiple contextual factors associated with suicidal behavior among adolescents attending two urban junior high schools in Indonesia using a socio‐ecological framework. Overall, 7% of participants were identified as being at high risk for suicide, with female students reporting a substantially higher prevalence than male students (11% vs. 3%). The overall prevalence observed in this study exceeds estimates reported in previous studies of Indonesian adolescents, in which 4% (Peltzer et al. 2017) and 3% (Putra et al. 2021) of students were identified as being at risk for suicide. This finding suggests that suicidal behavior remains an important public health concern and highlights the need to better understand the individual, interpersonal, and school‐related factors that may contribute to suicide risk among adolescents.

The association between suicidal behavior and student‐parent relationship variables observed in the present study is consistent with previous literature (Olcoń et al. 2017; Peltzer et al. 2017; Madjar et al. 2018; Gençöz and Or 2006). Specifically, stronger parent‐adolescent relationships were associated with a lower risk of suicide. This finding suggests that positive parent‐child relationships may foster positive and mutual communication, enabling adolescents to openly express their emotions (Kapetanovic and Skoog 2021), share their thoughts, and reduce the likelihood of interpersonal conflict (Choe et al. 2023). This type of supportive environment allows parents to intervene early if suicidal thoughts emerge. Parental support is positively associated with greater family cohesion (Harris and Molock 2000) and, along with the support of teachers and community leaders, may contribute to a safe and welcoming environment for adolescents (Olcoń et al. 2017). These supportive relationships may promote higher levels of emotional resilience and more adaptive coping strategies in the face of stressors and adversity occurring both within the family and broader social environment.

It is worth noting that the findings in the present study are not consistent with previous studies examining the association between parental supervision and suicide risks, although the results in this study approached statistical significance. One possible explanation could be that supervision may be more strongly associated with suicide risk among adolescents exhibiting delinquent or other problematic behaviors (Rohde et al. 1997), which was not specifically assessed in the present study. Alternatively, differences in the definition and measurement of parental supervision, variations in student's characteristics, or methodological differences across studies could also have impacted the results.

In addition to family‐related influences, the present study highlights the importance of the school environment, specifically teacher and peer support, as significant factors associated with suicide risk among adolescents. These findings underscore the protective role of social support with the school context and align with contemporary theories of suicidal behavior, including the IPTS, which emphasizes the importance of thwarted belongingness, and the Three‐Step Theory of Suicide (3ST), which identifies connectedness as a key protective factor. It also supports previous research conducted in the Middle East (Madjar et al. 2018), where adolescents reported that perceived levels of acceptance and caring from teachers and peers resulted in lower levels of suicidal ideation. A positive school climate—characterized by supportive teacher‐student relationships, inclusive peer interactions, and a sense of acceptance and social belonging—can play a critical role in preventing suicide among adolescents. This finding emphasizes the importance of teacher support in reducing the feelings of social disconnection, while peer support increases students’ self‐esteem and social bonds (Sun et al. 2006).

The present study also showed a strong association between bullying victimization and suicide risk, consistent with a number of studies demonstrating a significant relationship between these variables (i.e., Peltzer et al. 2017; Marthoenis and Yasir Arafat 2022; Klomek et al. 2009). The findings here highlight the relationship between peer victimization and suicide attempts, and suggests the strengthening of school‐based anti‐bullying interventions to mitigate the latent psychological consequences of peer bullying. In addition, future research should also examine potential mediating factors, including depression, which can influence the pathway between bullying victimization and suicide (Hertz et al. 2013).

Although the prevalence of illegal drug use (less than 1%) and alcohol consumption (1%) was relatively low in the present study, both behaviors were significantly associated with suicide risk. In contrast, smoking and vaping—reported by 9% and 8% of participants, respectively—were not significantly associated with suicidal behavior. One possible explanation is that smoking and vaping are relatively common among adolescents and the general population in Indonesia and have become common forms of experimentation during adolescence (Dishion et al. 1995; Tsering and Pal 2009). This widespread occurrence might reduce the perceived severity or psychological impact of these behaviors, making them less directly linked to suicidal ideation or attempts. Furthermore, substance use was not investigated in a deeper context (like whether the use was a one‐time experience or a regular occurrence driven by peer pressure), and this could account for the lack of a clear association. If adolescents engage in substance use just once or twice, possibly due to social influences like peer pressure, it may not have the same long‐term psychological consequences as more chronic or intense use, which could be linked to mental health issues such as suicidal ideation. Future research should include more specific measures of substance use patterns to better characterize these relationships.

The present study's logistic regression model highlighted the central role of parental relationships in reducing suicide risk, suggesting that positive parental involvement may support adolescents in managing their challenges and mental health conditions, a finding particularly salient within Eastern cultural contexts. Consistent with previous studies (Vermeiren et al. 2002; Peltzer et al. 2017; Putra et al. 2021; Olcoń et al. 2017; Geoffroy et al. 2016), exposure to physical aggression was also significantly associated with elevated suicide risk. Such victimization has been linked to emotional distress, including feelings of helplessness, humiliation, and intense psychological and physical pain, which may over time be associated with diminished self‐worth and depressive symptoms. In this regard, the literature suggests that depression may serve as a mediating factor in the relationship between physical aggression and suicidal behavior (i.e., Kiss et al. 2015).

One of the most profound findings in this study was the substantially higher risk of suicidal behavior among female students. Female adolescents were significantly more likely than male students to be classified as being at high risk for suicidal behavior, with females accounting for the majority of students identified as high risk. This finding is consistent with a growing body of literature demonstrating that adolescent girls report higher rates of suicidal ideation and suicide attempts than boys (Peltzer et al. 2017; Jeong et al. 2020; Geoffroy et al. 2016; Marthoenis and Yasir Arafat 2022; Olcoń et al. 2017; Putra et al. 2021). Additionally, as shown in Table 4, higher levels of truancy were associated with significantly increased odds in suicide attempt, suggesting that school disengagement may represent an important behavioral indicator of psychological distress among adolescents.

Table 4.

Correlation between independent variables with suicidal behavior.

Measure	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15
1. Suicidal behavior	—
2. Well‐spent time with family	−0.23***	—
3. Parental support of academic achievement	−0.17***	0.57***	—
4. Teacher support	−0.14**	0.22***	0.21***	—
5. Positive Peer support	−0.15**	0.25***	0.30***	0.32***	—
6. Student‐teacher perceived interaction availability	−0.11*	0.33***	0.35***	0.32***	0.37***	—
7. Exposure to a physical fight	0.12*	−0.04	0.01	−0.09	−0.05	0.00	—
8. Being threatened	0.15**	−0.03	−0.10*	−0.13**	−0.08	−0.01	0.22***	—
9. Experience of being physically attacked	0.20***	−0.08	−0.03	−0.05	−0.07	−0.01	0.28***	0.16**	—
10. Dating violence—physical abuse	0.13*	0.02	0.03	−0.12**	−0.08	0.04	0.22***	0.33***	0.20***	—
11. Dating violence—forced to have sexual intercourse	0.22***	−0.12**	−0.03	−0.06	−0.11*	−0.08	0.13**	0.21***	0.12*	0.11*	—
12. Truancy	0.23***	−0.09	−0.06	−0.07	−0.12*	−0.06	0.15**	0.14**	0.11*	0.11*	0.10*	—
13. Bullying victimization	0.25***	−0.08	0.04	−0.15**	−0.25***	−0.07	0.18***	0.05	0.32***	0.19***	0.14**	0.15**	—
14. Alcohol use	0.19***	0.04	0.00	−0.15**	−0.13**	−0.10*	0.05	−0.01	0.09	0.17***	0.23***	0.07	0.10*	—
15. Illegal drug use	0.17***	0.09	0.04	−0.10*	−0.15**	−0.03	0.13*	−0.01	0.09	0.31***	−0.01	0.15**	0.10*	0.58***	—
16. Prescription drug misuse	0.10*	0.02	0.01	0.00	−0.00	0.06	0.07	−0.03	0.05	−0.02	0.23***	0.00	0.10*	0.16**	−0.01

^* p < 0.05

^** p < 0.01

^*** p < 0.001.

Although experiences of dating violence involving forced sexual intercourse were not significantly associated with suicidal behavior, the magnitude of the observed association warrants careful consideration. Adolescents reporting these experiences had more than three times the odds of suicidal behavior than those who did not. The absence of statistical significance may reflect the low prevalence of reporting and the potential underreporting of experiences, particularly due to feelings of shame and self‐blame. Unintended (child) pregnancy among female students is universally stigmatized across all Indonesian cultures. Consistent with Epstein et al. (2020), this experience of forced sexual intercourse may prompt survivors to seek social invisibility and withdraw from school life, in an attempt to avoid perpetrators who might be encountered in the school environment. These results warrant further investigation due to the serious implications of dating violence on adolescent mental health in this specific urban context.

Several limitations should be considered when interpreting the findings of the present study. First, the absence of an association between perceived school safety and suicidal behavior contrasts with previous literature (Nickerson and Slater 2009; Olcoń et al. 2017). This may reflect residual post‐COVID‐19 influence, whereby perceptions of school safety may have been influenced by concerns related to the pandemic rather than by the school environment itself. Additional research is needed to examine this possibility.

Second, the original three‐category suicide risk variable (no risk, medium risk, and high risk) was recoded into a binary outcome (no risk vs. risk) for the risk modeling analyses because the relatively low prevalence of medium‐ and high‐risk cases resulted in sparse data that could have produced unstable parameter estimates (Peduzzi et al. 1996). Moreover, the primary objective of this study was to support population‐level screening and estimate the prevalence of suicide risk rather than to distinguish between different levels or clinical profiles of suicidal risk.

Third, predictors with very low base rates (1%–3%) including “being threatened,” “forced to have sexual intercourse,” and “prescription drug misuse” produced wide confidence intervals and unstable estimates, possibly compounded by underreporting of sensitive items (Tourangeau and Yan 2007). Finally, the exclusion of 106 participants resulted in a reduced sample (n = 446 vs. required n = 567), which may have limited statistical power to detect additional significant predictors. Nonetheless, comparisons between excluded and retained groups revealed no significant differences in age, gender, grade level, or 10 of 13 regression variables, with Cronbach's alpha (0.72 vs. 0.79), suggesting that exclusion bias was unlikely.

An additional limitation relates to the survey instrument itself. The present study utilized data from an adapted version of the SCSS School Climate Survey‐Student version, which contains a large number of items. Though the “normative” completion duration was utilized to exclude careless responses in the analysis, there was still a chance that respondent fatigue and insufficient effort in responding threatened the true variance. Future research should consider using shorter questionnaires or simplifying the questionnaire structure as well as their interface in the online platform, while maintaining adequate measurement of key constructs.

5. Conclusion

The present study contributes to the limited body of research examining suicide risk among adolescents in an urban Indonesian context by providing evidence on both the prevalence of suicidal behavior and the contextual factors associated with elevated risk. Guided by a socio‐ecological framework, the findings demonstrate that suicide risk is shaped by multiple influences across the individual, family, and school environments. In particular, the study identifies several factors associated with increased suicide risk while highlighting the importance of supportive family relationships and positive school environments as potential protective influences.

These findings have several implications for adolescent suicide prevention strategies, in particular in Indonesia. First, family‐based interventions that strengthen parent–adolescent communication, emotional support, and family cohesion may be particularly valuable and should be considered as a key component in comprehensive suicide prevention strategies. At the school level, prevention and intervention efforts should prioritize students at elevated risk, including adolescents who are female, students who are frequently truant, and those who are exposed to physical and sexual violence. It is encouraged for the schools to play a critical role in these efforts by implementing gatekeeper training programs that equip teachers and school staff to recognize and respond to early warning signs of suicide risk, including social withdrawal, persistent absenteeism, and victimization. Additionally, strengthening anti‐bullying policies and cultivating supportive peer environments may help reduce key risk factors associated with suicidal behavior. Together, these strategies may contribute to a more comprehensive and contextually appropriate approach to adolescent suicide prevention in the Indonesian context, while also offering insights that may be applicable to adolescents living in similar community settings in other rapidly urbanizing regions.

Author Contributions

Conceptualization: Dicky C. Pelupessy. Methodology: Dicky C. Pelupessy, Dimas B. Prasetyo, Susanne Argamaso, and Sabrina Arredondo Mattson. Formal analysis: Dimas B. Prasetyo, Amelia Putri, Dicky C. Pelupessy, Susanne Argamaso, Sabrina Arredondo Mattson, Shr‐Jie Wang, and Beverly Kingston. Writing, review and editing: Dicky C. Pelupessy, Dimas B. Prasetyo, Amelia Putri, Ferena Debineva, Susanne Argamaso, Sabrina Arredondo Mattson, Shr‐Jie Wang, and Beverly Kingston. Project administration: Ferena Debineva and Amelia Putri.

Disclosure

All authors read and approved the final manuscript.

Ethics Statement

This study has been approved by the Ethics Committee of the Faculty of Psychology, Universitas Indonesia (No. 165/FPsi.Komite Etik/PDP.04.00/2022).

Consent

Participants of the study provided their informed consent before participating in the survey.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions. The data used can be requested from the authors.