The effects of conflict on adolescent girls and young women: intersecting forms of violence

Globally, approximately 89.2 million adolescent girls aged between 10 and 17 years live in conflict-affected environments, representing one fifth of all adolescent girls worldwide [11]. Accurate data on conflict-exposed adolescents in Myanmar have not been collected, but according to the United Nations Office for the Coordination of Humanitarian Affairs [7], nearly 40% of internally displaced people in Myanmar are children. Repeatedly, studies demonstrate that adolescent girls and young women are disproportionately at risk of the severe direct and indirect health effects that occur during and after conflict, but they are often overlooked in international humanitarian responses [12].

As well as direct harm from combat, girls in conflict-affected settings are exposed to increased risks of sexual and gender-based violence, including forced marriage [13,14]. An estimated 39% of women and girls in conflict-affected settings have been exposed to physical or lifetime intimate partner violence according to a recent meta-analysis, and in 2023, 3,688 acts of conflict-related sexual violence were confirmed by the UN, though this is widely considered a significant undercount [15,16]. Collectively, studies show that exposure to sexual violence causes physical, as well as psychological, trauma, and a heightened risk of the likelihood of sexually transmitted infection and unwanted pregnancies [17]. For instance, globally, girls in conflict-affected settings are 20% more likely to marry as children than girls in non-conflict areas [11]. High rates of post-traumatic stress disorder (PTSD), depression, and anxiety have also been documented among adolescent girls in conflict areas, further compounded by stigma over sexual violence [18].

Concurrently, conflict also impedes critical health services, limiting access to sexual and reproductive healthcare, education and psychosocial services [19]. Economic uncertainty and school closures further limit adolescents’ education and future job prospects [20]. Without these protective systems, adolescent girls are at increased risk of human trafficking, communicable disease, and social isolation, which, together with the erosion of essential services, can reinforce cycles of poverty and accelerate poor wellbeing and health outcomes [13,21].

In view of significant research into violence against women and girls in conflict settings, the unique vulnerabilities among adolescents require further exploration. Often, studies in conflict settings—and in general—fall short of addressing the intersecting risks faced by adolescents, who may be exposed to both intimate partner violence (IPV) as well as violence against children (VAC) [22]. IPV refers to physical, sexual or psychological abuse by a current or former partner and is often researched with respect to adult relationships [23]. On the contrary, VAC covers all forms physical and psychological violence, neglect, maltreatment, and exploitation including sexual abuse by either family or non-family members [23]. Adolescents aged between 12–18 years, however, exist at the intersection of these two types of violence such that they can experience IPV in early relationships whilst still being at risk of VAC by parents, caregivers or other adult persons [22].

Specific studies on VAC by parents towards their adolescent children are rare, but broader evidence including younger children indicates an association between conflict exposure and heightened parent-perpetrated VAC. For instance, Malcom et al. [24] reported that parents in conflict-affected areas versus non-conflict areas of Iraq were significantly more likely to use moderate and severe methods of corporal punishment. Similarly, research findings suggest associations between conflict exposure and more parental aggression in Palestine [25], child sexual abuse in Lebanon [26] and child maltreatment in Sri Lanka and Uganda [27]. Similarly, various studies in the United States have indicated that child maltreatment is elevated within families with one or both parents being members of the US military [14,28–30]. For example, Hisle-Gorman et al. [31] demonstrated that parental deployment was associated with child maltreatment against children aged three to eight, and for children of injured veterans in particular, there was a 24% rise in child maltreatment with each additional parental injury. Likewise, a meta-analysis carried out by Eltanamly et al. [32] revealed that war-exposed parents treated their children harshly and were less warm. That said, there is not a straightforward relationship between conflict exposure and parenting. Not all families with parents exposed to armed conflict respond with harshness at home. Various individual, familial and societal protective factors can reduce violence, which highlights the need for additional context-sensitive and nuanced research in this area.

Several theories help explain why exposure to armed conflict can escalate violence at home. Murphy et al. [33] synthesised published research and theoretical perspectives to create a conceptual framework for the drivers of violence against women and girls in conflict and post conflict settings. Their findings document risk factors for violence in six spheres: global, societal, institutional, community, interpersonal, and individual. Similarly, qualitative research from the Democratic Republic of Congo and Myanmar used an ecological framework approach to map out a network of intersecting risk factors and social norms that contribute to domestic violence in humanitarian settings [34]. Both studies highlight gender inequality as a key macro-level driver, which permeates into community and household dynamics, where conflict exacerbates hyper-masculine and patriarchal norms, weakens the rule of law, and normalises violence. Meanwhile, household stress—exacerbated by financial instability, food insecurity, and personal insecurity—increases the risk of domestic violence, particularly when individuals turn to negative coping mechanisms such as alcohol and substance abuse [33–35]. Although these theories focus on women and girls more broadly, they are relevant for adolescent experiences of violence, because of the shared risk factors between IPV and violence against children, and adolescents’ risk of each form of violence [36].

There is a lack of research on parent or caregiver violence towards adolescents in Myanmar. A study by Aye et al. [37] surveyed 2,377 adults and found that 21.1% reported experiencing abuse during childhood, more commonly reported by women (29.8%) than men (12.4%). According to the 2016 Global School-Based Student Health Survey, 32.7% of students aged 13–17 years in Myanmar said they had been physically attacked in the past year, more common among boys (39.8%) than girls (26.3%) [38]. In another study, Nyan Linn et al. [39] analysed data from the 2015–2016 Myanmar Demographic and Health Survey (DHS) and found that 44.5% of children aged 2–14 years had experienced corporal punishment, again with a higher prevalence among boys than girls. A study by Miedema & Kyaw [40] based on the same 2015–2016 DHS data identified that, among children ages 2–14 years, 18% reported abuse from a female caregiver, which was associated with women’s past year exposure to physical or sexual IPV. Despite this, formal systems for preventing or responding to VAC in Myanmar are scarce and policy legislation around physical discipline of children remains ambiguous [41]. Although Myanmar adopted a new Child Rights Law in 2019 prohibiting all forms of physical punishment of children, such acts are still widely accepted in the home, day care, schools and alternative care settings [41,42]. Ultimately, efforts to prevent or address VAC in Myanmar remain limited, and research is needed on how the ongoing conflict may influence the prevalence of VAC, especially among adolescent girls, who face intersecting risks related to their age and gender.

Gendered parenting practices and violence

“Gendered parenting”, also known as differential parental socialisation, describes the different ways in which parents may treat their children based on their gender [43,44]. For instance, in focus group discussions with adolescent girls and young women in Myanmar, participants described how their parents were harsher and placed more restrictions on their freedoms than they did to their brothers [45]. The participants described how this unequal parenting had worsened since the 2021 coup, perhaps driven by factors such as heightened parental concern for their safety and growing financial stress at home. These gendered approaches to parenting are often rooted in beliefs about how boys and girls should behave, and they tend to reflect and reinforce inequitable gender norms [46]. For example, qualitative research from Uganda found that boys experienced greater freedom to play in their free time when compared to girls, who instead reported spending their free time learning gender-specific roles such as cooking [47]. Such practices not only affect childhood experiences but can also shape attitudes and behaviours that carry into adulthood.

Although gendered parenting has received limited attention in conflict-related research, it warrants a greater focus because of its potential role in helping us to better understand violence towards young women. Gendered parenting practices can lead to long-term disempowerment of daughters by reinforcing male dominance and traditional gender stereotypes [48]. Understanding this is especially important for conflict settings, where parents or caregivers (often hoping to protect their daughters’ safety and reputations) may restrict their daughters’ autonomy, mobility, and sometimes facilitate early or forced marriage [49]. The impacts of early and forced marriages, which include unwanted pregnancy, increased IPV and poorer mental health, are particularly harmful in humanitarian crises, where access to health care and support services is critically limited [50–52]. Evidence from conflict-affected settings, such as the Democratic Republic of Congo, suggests that caregiver’s gender-equitable attitudes was associated with reduced risk of sexual abuse among adolescent girls, and reduced likelihoods of girls’ acceptance of IPV [53]. In Myanmar, reports suggest that early and forced marriages have increased since the introduction of the 2024 forced conscription law, which exempts married women from military service [54].

Gender inequality and societal norms

Gender inequality is prevalent across Myanmar, stemming from historically patriarchal traditions and societal norms that shape the expectations of girls and women from a young age. A key example is the concept of hpon, a widespread notion of male superiority over women under Theravada Buddhism in Myanmar [55]. Hpon perpetuates beliefs around women’s impurity, positioning this impurity as a direct threat to men’s hpon, which has contributed to women being excluded from leadership roles and restricted from certain religious spaces [55,56]. Similar gendered hierarchies exist among other religious and ethnic communities, and research has shown that women from minority backgrounds are disproportionately impacted by these inequalities [56]. Women and girls across the country risk stigma and consequences if they fail to uphold the gendered expectations of obedience and modesty that are expected of them, and adolescent girls are especially impacted by these gendered norms due to their young age [55–57]. For example, from an early age, many girls are socialised to take on household responsibilities and gendered roles within the family to be considered a “good daughter”, and—despite greater emphasis on girls’ education in recent years—their schooling is still widely considered to be less important than boys’, especially in financially insecure settings [58]. Furthermore, girls must navigate entrenched taboos around their sexuality and reproductive health and can face stigma following experiences of gender-based violence, restricting their ability to access information and support and putting their health and wellbeing at risk. These dynamics are further heightened during times of conflict, economic instability and displacement, which simultaneously exacerbate risk to girls’ health whilst also weakening the systems that may otherwise protect them [58–60].

Conceptual framework, aims and objectives

Fig 1 outlines the conceptual framework for this analysis, based on our previous qualitative research [45] and our review of the published literature. Fig 1 illustrates how the situation of armed conflict is reinforcing an existing system of gender inequality and societal norms. We believe that both the armed conflict and structural gender inequalities shape gendered parenting practices and parent/caregiver perpetrated violence. We also believe that gendered parenting practices may increase the prevalence of parent/caregiver perpetrated violence. Ultimately, we predict that both gendered parenting practices and parent/caregiver violence likely result in negative mental health and wellbeing outcomes for adolescent girls and young women. These health and wellbeing outcomes will be at greater risk due to the continued political instability, which is blocking the already-limited mental health and psychosocial support services, especially in the rural and conflict-affected areas [61].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Conceptual framework for our analysis, summarising hypothesised links between armed conflict, gender inequality, gendered parenting practices and parent/caregiver perpetrated violence among adolescent girls and young women.

The purple boxes represent domains shaped by both the armed conflict and gender inequality, to reflect their compounding influence.** highlights the data informing this study.

https://doi.org/10.1371/journal.pgph.0004737.g001

Despite the substantial risks of harm to young women and girls during conflict, to our knowledge, few studies have explicitly investigated the association between gendered parenting and violence against adolescent girls, and there have been none that consider this amidst the conflict in Myanmar. This study aims to measure the prevalence of violence by parents or caregivers among a sample of adolescent girls in Myanmar and examine associations between parents’ use of violence and girls’ experiences of gendered parenting following the 2021 military coup.

Ethics Statement

For participants under the age of 18 years, we collected written informed assent from all participants and either written informed parental consent or the written informed consent of a trusted adult on behalf of each participant. For participants aged 18 years or over, informed written consent was collected. Ethical approval for this research was received from the London School of Hygiene and Tropical Medicine Ethics Committee (ref 28458–2) and a local research ethics committee in Myanmar. Before enquiring about sensitive topics, participants were reminded of the option to skip any questions. All interviews were conducted in enclosed rooms where privacy could be maintained and all peer-researchers had prior experience researching sensitive topics, particularly violence. Breaks were offered throughout the course of the interview and participants were informed that they were able to withdraw their consent and end the interview at any time. Participation was voluntary and no monetary compensation was provided as the interviews were conducted as part of participants’ ongoing GD programme schedule. Referral pathways were ensured through our partner organisation, GD.

Inclusivity and global research

Additional information regarding the ethical, cultural, and scientific considerations specific to inclusivity in global research is included in the Supporting Information (S1 Checklist)

Study design and participants

We conducted this study in collaboration with Myanmar-based NGO, Girl Determined (GD). GD supports girls and young women aged 11–25 years from low-income rural and urban settings across Myanmar. The two-year core programme, Circles, is an in-person weekly girls empowerment programme, facilitated by women hired and trained from local communities. Circles covers topics such as: self-confidence, assertiveness, ethical and religious discrimination, trafficking and safe migration practices, stress management, relationships (both romantic and friendships), drug awareness, goal-setting and sexual and reproductive health education. GD recruits hard-to-reach girls through sites such as camps for internally displaced people, churches, Buddhist nunneries and dormitories for girls from rural areas. Therefore, all girls enrolled in Circles generally represent girls from the lowest socio-economic quintiles in the regions of Myanmar that GD is active in. Due to GD’s targeted outreach to nunneries and boarding houses, our sample included a high proportion of girls who were living away from home and enrolled in education, as these institutions provide education and accommodation to girls from low-income communities across Myanmar. At the time of survey, each participant was enrolled in GD’s Circles programme and enrolment occurs at the site-level, for example, all girls aged 10–18 years who are residing at a selected boarding house will be invited to join. For more details about GD’s work, please see Girl Determined [58].

The design and content of the survey was co-developed with six young, female peer-researchers from GD (EES, HJ, KL, NDO, NMS, ZMH), four of whom (EES, KL, NDO, ZMH) also co-developed and facilitated a prior qualitative study which helped inform the conceptual framework for this study see Pearson et al. [45] for additional details. The peer-researchers conducted the surveys in person, after receiving training on data collection methods and research ethics from a British doctoral student (IP) and a Myanmar research assistant and translator CVK. Five sites were chosen for data collection, which were the home states of the peer-researchers and thus safe and accessible to them during the data collection period. The sites were: Yangon Region, Mandalay Region, Kachin State, Shan State and Tanintharyi Region—all sites where GD implements their empowerment, digital literacy and sports programmes. Within each region, we randomly selected sites where GD were already implementing the Circles programme. We randomly selected sufficient sites for each peer-researcher to conduct 125 surveys, whilst purposely excluding any that were unsafe due to conflict or weather. Two peer-researchers were based in Mandalay Region and one each in Yangon Region, Shan state, and Kachin Region and one peer-researcher conducted interviews across both Yangon and Tanintharyi Regions. The participant recruitment period lasted from 22/09/2023 until 16/01/2024. At each site, we invited every girl registered with GD (most of whom are aged 10–18 years) to complete a survey. The main reason girls declined to take part in the study was school exams.

Procedures

The survey tool was co-developed with the peer-research team in September 2023 in Yangon which involved multiple participatory workshops where peer-researchers reviewed, refined and contributed to the adaptation of survey questions to ensure cultural and contextual appropriateness. Most survey questions are based on existing tools, some were also validated for use with adolescents in low-income settings. None of the questions we used were validated in our setting due to the scarcity of existing Myanmar-validated tools and our own time and resource constraints. During our co-production workshops, our survey tool was reviewed and developed by the research team to ensure it was culturally and contextually appropriate. CVK translated the survey tool, which was checked by bilingual members of GD staff. We conducted eight pilot interviews, and further developed the tool based on the pilot-results. For questions with Likert scale responses, a card was presented to participants with a visual representation of the scale using smiling faces and a green to red colour code.

The survey was conducted between September 2023 and January 2024. The peer-researchers read all questions out loud to the participants and then recorded their verbal responses into Kobo Toolbox using mobile phones. Each survey lasted approximately one hour and collected data on participant demographics, their living situation, social support, exposure to conflict, experiences of parent or caregiver perpetrated physical and psychological violence, mental health and hope. An in-depth analysis of the results pertaining to mental health and conflict exposure have been published elsewhere [62]

Measures

Statistical analyses

We summarised key demographic and behavioural variables using descriptive statistics, stratified by geographical region and age group. We measured exposure to physical and psychological violence in the past year using binary indicators derived from participant responses to the violence-related questions. An affirmative response to any question categorised participants as exposed. For the gendered parenting scale, we coded responses numerically (−2 to +2, where disagreement indicated lower gendered parenting bias) and summed to generate a continuous gendered parenting score. Scores were summarised using means, medians, standard deviations, and ranges. We conducted binomial tests to compare agreement versus disagreement for each question. We assessed conflict exposure using a categorical variable representing three levels of exposure (none, 1–2 events, 3 + events), with dummy variables created for logistic regression analyses. We assessed associations between violence and age group, as well as living arrangements, using X² tests and proportions for each subgroup. To analyse the association of gendered parenting and conflict exposure on violent parenting, we ran two logistic regression models: one for physical violence and one for psychological violence. Predictors included conflict exposure (dummy-coded), gendered parenting score, age, and whether participants lived with their birth parents. We calculated adjusted odds ratios (aORs) and 95% confidence intervals (CIs) to quantify associations. We conducted analyses in R Studio (version 4.3.0).

Demographics

Table 1 presents the demographics of our study sample. Study participants included 731 adolescent girls and young women, aged 10–17 years. The overall median age was 14 years. Participants were living in villages and peri-urban areas in Mandalay (34.1%) and Yangon Regions (26.4%), as well as Kachin State (16.6%), Shan State (17.1%) and Tanintharyi Region (5.9%). Approximately half of participants (51.4%) were living away from their parents at the time of the survey, which was more common among mid-late adolescents (67.8%) than early adolescents (42.2%). Those living away from their parents were most often living at Buddhist nunneries (39.5%). Overall, 98.1% of our participants were enrolled in some form of education at the time of the survey.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Participant demographics.

https://doi.org/10.1371/journal.pgph.0004737.t001

Violence perpetrated by a parent or caregiver

A total of 715 participants chose to answer the questions on their experience of parent/caregiver-perpetrated psychological and physical violence, hereafter referred to as caregiver-perpetrated violence. The participants most often listed parents as their primary caregivers. Non-parent caregivers were primarily other family members but also included Nuns and boarding house teachers for some of the girls who were living away from home. Fig 2 depicts an overview of the results and Table 2 provides a detailed breakdown of the different physical and psychological violence types and their reported frequency.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Prevalence of physical and psychological violence types (n = 715) also be noted that some of the girls who lived away from family members still listed parents as their primary caregiver in the survey.

https://doi.org/10.1371/journal.pgph.0004737.t002

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Graph to show percentage of participants reporting different violence types (n = 715).

https://doi.org/10.1371/journal.pgph.0004737.g002

Of the 715 participants, 80.6% experienced at least one form of psychological violence in the past year and 49.8% experienced at least one form of physical violence. Both past-year physical and psychological violence were more frequently reported by the early adolescents, but this difference was only statistically significant for physical violence (53.1% of early-age adolescents vs. 44.7% of mid-to-late age adolescents; X² = 4.3, p = 0.038). For psychological violence, there was no significant difference between age groups (83.4% early adolescents vs. 80.6% of mid-late adolescents, X² = 0.68, p = 0.411). The most frequently reported type of psychological violence was being shouted, screamed or yelled at loudly and the most frequent physical violence was to be hit with an object (Fig 2).

Caregiver perpetrated psychological violence was reported more often by participants who lived either with their parents or another family member (51.5%) versus those living away from family (48.5%). However, this difference was not significant (X² = 3.1, p = 0.076). Conversely, caregiver-perpetrated physical violence in the past year was more often reported by participants who lived away from family members (51.1%) compared to those living with family members (48.9%), however, this difference was not statistically significant (X² = 0.38, p = 0.540). It should

Gendered Parenting

A total of 514 participants who had brothers chose to respond to the questions about gendered parenting. Fig 3 provides a summary of the percentage of participants who agreed or disagreed with the eight statements, and more detailed results are presented in Table 3. The mean gendered parenting score was -0.95 (SD = 7.11). Since this score is close to zero, it suggests that participants generally gave neutral responses, with a slight tendency toward disagreeing with statements that their parents showed favouritism to their brothers.

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Results of a binomial test to compare the proportion who agree with those who disagree for each of the gendered parenting questions (n = 514).

https://doi.org/10.1371/journal.pgph.0004737.t003

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Percentage of participants (n = 514) who agreed or disagreed with statements comparing their experiences of gendered parenting to how they perceive their brother(s) to be parented.

Agreement (right-hand side) and disagreement (left-hand side) are shown across four response categories: strongly disagree, slightly disagree, slightly agree, strongly agree.

https://doi.org/10.1371/journal.pgph.0004737.g003

Table 3 also includes the results of binomial tests conducted for each of the eight gendered parenting statements. These tests measured whether significantly more participants agreed or disagreed with each statement. The results show that participants were significantly more likely to agree with certain statements suggesting unequal treatment compared to their brothers. For example, 74.3% agreed that they had less freedom to see friends, 68.3% said they received less praise for similar tasks and 62.8% said they were assigned more chores, all of which were statistically significant (p < 0.001). On the other hand, participants were significantly more likely to disagree with statements suggesting that they had to marry earlier than boys (85.8% disagree, p < 0.001) or drop out of school earlier (78.4% disagree, p < 0.001). For statements about receiving lower allowances, being trusted less, or feeling less loved, there were no significant differences (p > 0.05).

Association between perceived gendered parenting, exposure to armed conflict and caregiver-perpetrated violence

The prevalence of conflict exposure and the types of conflict that the study population were exposed to have been detailed elsewhere [62]. In the present study, we conducted a multiple logistic regression analysis to test for associations between the number of conflict exposures and gendered parenting scores on caregiver-perpetrated violence perpetrated in the past year, whilst also accounting for participants’ age and living status. The results are presented in Table 4.

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. Multiple logistic regression results for past year parent and caregiver perpetrated physical and psychological violence (n = 514).

https://doi.org/10.1371/journal.pgph.0004737.t004

The analysis presented in Table 4 found that participants who reported exposure to three or more conflict-related stressors were twice as likely to report physical violence in the past year compared to those without any conflict exposure (adjusted OR=2.19, 95% CI = 1.24-3.89, p = 0.007). The analysis also found that participants who scored higher on the gendered parenting scale were more likely to report physical violence, as each additional point on the scale was associated with a small but significant increase in the likelihood of experiencing violence (adjusted OR=1.06, 95% CI = 1.03-1.09, p < 0.001). Younger participants were also more likely to report physical violence, with the odds decreasing as age increased (adjusted OR=0.83, 95% CI = 0.73-0.95, p = 0.007). However, whether or not a participant lived with their birth parents was not significantly associated with reporting physical violence (p = 0.879).

Participants who reported exposure to 1–2 conflict-related stressors were twice as likely to report psychological violence in the past year when compared to those who did not report any exposure to conflict (adjusted OR=2.04, 95% CI = 1.09-3.77, p = 0.025). In addition, higher scores on the gendered parenting scale were also associated with an increased risk of psychological violence. While the increase per point was relatively small, it remained statistically significant (adjusted OR=1.07, 95% CI = 1.03-1.10, p < 0.001). By contrast, there was no significant relationship between psychological violence and either the participant’s age (p = 0.976) or whether they were living with their birth parents (p = 0.274).

Benefits and limitations

This study offers a novel insight into the lives of adolescent girls in conflict-affected Myanmar and provides a unique analysis of the intersections of conflict, violence and gendered parenting practices, which to the best of our knowledge has not been investigated before in this setting. Furthermore, this data provides important evidence for just some of the many negative impacts that Myanmar’s protracted conflict is having on adolescent girls, and in doing so advocates for an often-overlooked demographic. That said, besides the specific limitations related to the use of proxy measures, there are a few additional considerations when interpreting the results of this study. First, as mentioned, none of the survey tools were officially validated with adolescent girls in this setting. However, our co-production approach and in-depth co-production workshops—where we dedicated time to developing the survey tools in collaboration with young peer-researchers who were previously among the study population demographic themselves—significantly increased our confidence in the tools. We believe that although not validated in our setting, our co-production approach allowed for comprehensive tool-development given our time and resource constraints.

For the violence-related questions especially, we relied on self-reported data, which are at risk of recall bias. In particular, telescoping bias, where participants may recall events as occurring more recently or further in the past than they actually occurred [77]. Furthermore, studies have shown that participants with post-traumatic stress disorder show bias towards recalling negative information which could be relevant to our study population [78]. To mitigate this recall bias, we limited our violence analysis to only consider violent acts that happened in the previous year. However, we did not assess the severity of the violence experienced, meaning that all forms of violence were weighted equally in our analyses, regardless of the type of act. This approach may have overlooked important distinctions between different types of violence and their relative severity. Additionally, as this survey was conducted in 2023, we allowed for a two-year window of conflict exposures since the 2021 military coup, meaning that the conflict data is at slightly higher risk of recall bias [79]. Finally, our measure of parent/caregiver-perpetrated violence did not also enquire about demonstrations of love, support or care, meaning our measures of violence are being considered in isolation of any broader changes in parent/child relationships.

Our gendered parenting scale is limited in that we did not enquire with the parents or the brothers of our participants, meaning that our conclusions are based wholly on the subjective perceptions of the participants, introducing a range of self-reporting biases such as perception bias, recall bias, confirmation bias and social desirability bias. Due to the risk of multiple forms of bias in our gendered parenting measure, the limited generalisability of our findings should be noted, and we acknowledge that our findings may not fully capture the broader family dynamics or the interactions of parents and brothers. Further research into gendered parenting practices should include multiple perspectives and additional methods such as observations to help triangulate the results. Moreover, while we aimed to use random sampling, the ongoing insecurity and instability across Myanmar sometimes requires last-minute revision of survey sites, meaning random sampling was not always possible. Participants were invited who were current beneficiaries of GD’s empowerment programme, which limits the generalisability of findings to girls outside this programme. It should also be noted that the participants of this study had already received training and education on topics such as gender equality and violence, which may have influenced their responses to questions around these topics. Furthermore, the region-based sampling could have created clustering effects by region. Since we only included five regions with modest sample sizes in each, we did not use multilevel modelling or adjustments for clustering affects. Finally, as with all cross-sectional studies, we cannot investigate the causal pathways between our variables or make robust predictions about mediation or moderation. That said, we can conclude that there are statistical associations between conflict, violence and gendered parenting practices within our study population, and therefore future research into tailored violence-prevention interventions should take this into account.

Furthermore, while this study offers important insights into the quantitative associations between key variables, the complexity of adolescent girls’ experiences highlights the need for complementary qualitative research to better understand the nuances behind these intersecting risk factors. While outside the scope of this study, qualitative data from parents and caregivers would offer a further understanding of gendered parenting practices and the use of violence within the home, including why these behaviours may be exacerbated during periods of instability and uncertainty. Such research could also help to identify contextually appropriate targets for future interventions to protect adolescent girls from these risk factors.

Conclusion

Our findings highlight a significant overlap between armed conflict, gendered parenting practices, and parent or caregiver perpetrated violence. Given the documented risks of abuse and conflict on the health and wellbeing of adolescent girls, groups working in conflict-affected settings should be alerted to both exposures when considering health and other interventions. Specifically, the prevalence of parent/caregiver perpetrated violence and its strong association with conflict and parental gender norms demands responses in these settings that consider risks at multiple levels of the socioecological framework. The finding that a high prevalence of physical violence was found among younger adolescents in particular highlights the necessity for age-specific responses. Lastly, while armed conflict poses immediate and serious risks, our results highlight the potentially long-term consequences for girls and young women, impacting their long-term safety, autonomy and overall health and wellbeing. Interventions should ensure the physical safety and psychological wellbeing of girls and young women, whilst also engaging boys and young men to ensure they are not reflecting their parents’ or society’s negative gendered attitudes. These findings highlight the need for adaptable, multi-sectoral interventions that reflect the complexity of conflict-affected settings and girls’ wellbeing, which requires activities related to violence prevention, mental health support, empowerment, and the shifting of harmful gendered norms within families and communities. Programmes such as Circles offer a valuable opportunity to support girls and young women, not only directly, but also through targeted outreach to parents and caregivers. This study’s findings may serve as a useful advocacy tool to encourage caregivers to reflect on their parenting practices and to inform intervention strategies that promote safer and more gender-equitable home environments.