Study design and participants

The study included both direct contact violence and non-contact forms, while also examining the dynamics of the survivor-perpetrator relationship. We employed an adapted questionnaire [19] to conduct a comprehensive survey including 961 adolescents in the designated region. The survey was carried between 1^st May and 28^th August 2021.

Due to Nigeria’s size and ongoing political instability, nationally representative studies are challenging. Instead, we focused on the geo-political zone of the South-West, which includes the important economic centers of Lagos and Ibadan and is home to 22% of all Nigerians. Data provided by UNICEF indicates that approximately 92% of adolescents in the South West region of Nigeria complete their primary education, while approximately 85% proceed to complete their secondary education. This suggests that there is a significant proportion of adolescents in the region who are not completing their education, estimated to be between 8 and 15% [36]. A cross-sectional design was adopted in this study. Ekiti state was excluded due to the deteriorating public security situation. Both in- and out-of-school adolescents aged 13-17 were the target population.

In-school adolescents – a multistage stratified sampling method was adopted and data were collected in five South-West Nigerian States. 27 public secondary schools from rural and urban communities were randomly sampled: Oyo (= 5), Osun (= 6), Lagos (n = 5), Ogun (n = 6), and Ondo (n = 4). Based on the Ministry of Education’s list of registered public secondary schools in the South-West geo-political zone, we selected the sample by using generated numbers on each community (state, local government area, public secondary schools, and current level of study). In each school, a class teacher was assigned by the head teacher to assist in generating sample frames through the class registers from which we randomly selected the participants. Between 15 to 45 adolescents were interviewed (n = 862).

Out-of-school adolescents – Given the challenges of reaching adolescents who do not attend school, systematic sampling was not possible, thus we collected a convenience sample in street locations where these adolescents congregated. This included bus stations and markets where they sell small goods and carry loads for stallholders and customers, as well as where they sleep (abandoned buildings, under bridges) and at roadsides where charitable donations are provided. Due to this method, we reached a severely disadvantaged sample of adolescents (n = 99).

Procedures

Following established international guidelines [37], independent Yoruba native speakers, experienced in research and translation, translated all measures from English to Yoruba and then back to English. SurveyToGo software was used for the data collection, using individual tablets (except with subsample of the out-of-school, see Strengths and Limitations for details).

Ethical considerations

We followed ethical guidelines as outlined by Mathews et al. (2022) [35]. The protocol was reviewed and approved by the Institute for Advanced Medical Research and Training (IAMRT), College of Medicine, University of Ibadan (UI/UCH number: UI/EC/21/0031) and the Institutional Review Board (Ethics Committee) of the University of Konstanz (IRB number: 21/2021). The study was conducted in accordance with the declaration of Helsinki. Data was collected with tablets and anonymized. Verbal and written consent were obtained from participants involved in this study. During data collection, following permission from head teachers, we explained the purpose of the study in the classroom and provided the adolescents with a parental consent form. On the following day, depending on individual and a signed parental consent, we carried out the survey. For out-of-school participants we could not obtain parental consent, because most do not live with their parents neither had a legal guardian. Almost all of these adolescents organize their lives independently, so it seemed appropriate to rely on their informed consent.

Enumerators received extensive training on scientific procedures of data collection, resilience, risk assessment and ethical protocols such as the need to ensure confidentiality, safety and privacy of the respondents. Moreover, they were trained and instructed to refer participating adolescents exhibiting marked psychological distress or who reported ongoing emotional, physical, or sexual abuse to the study psychologist who conducted a follow-up interview. A total of 53 adolescents were referred to either school counsellors or NGOs.

Enumerators as well as participants strictly adhered to the Covid-19 Protocols by wearing face masks. Refreshments (snacks) and 200 Naira (about $0.5) were given to each participant after the assessment, they had no prior knowledge of this small compensation of time and efforts.

Due to the sensitive nature of the topic we used gender-matching techniques, i.e., only female enumerators were responsible for guiding and interviewing the female participants and only male enumerators attended to the male participants.

Participants characteristics

In a total of 961 participants in Nigeria, we asked every participant how they identified regarding gender and all opted for binary gender identification (518 girls and 443 boys). The sample comprised adolescents aged 13 to 17, with 95.9% aged between 15 and 17 years. Table 1 provides a comprehensive overview of participants’ characteristics. Urban communities constituted a majority (75.7%) of the surveyed areas, and a distinct difference emerged in living arrangements between in-school and out-of-school adolescents. While 88.6% of in-school adolescents lived with either one or both parents, a significant majority (56.6%) of out-of-school adolescents did not reside with their parents. Financial support primarily came from parents for in-school adolescents (91.6%), with a notable contrast in out-of-school adolescents (48.5%). Emotional support was predominantly derived from one or both parents (88.6%), but this was markedly higher among in-school participants (92.8%) compared to out-of-school respondents (51.5%). Regarding romantic relationships, 57.6% of participating adolescents reported being in one. This percentage was substantially higher among out-of-school adolescents (76.8%) in comparison to their in-school counterparts (55.5%). A noteworthy 20.1% of adolescents reported an age gap of five or more years with their partners. A substantial proportion (43.8%) of adolescents indicated experiencing some form of SV before the age of 12, with a marked difference observed between in-school (41.4%) and out-of-school (64.6%) adolescents. Additionally, a wealth index was constructed based on household assets, showing that while the average adolescent lived in a household possessing 5.74 assets, the out-of-school group had significantly fewer assets (2.53 assets). Moreover, a substantial difference emerged in the proportion of households owning seven or more assets (9.1% for out-of-school versus 31.0% for in-school adolescents).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Participants characteristics.

https://doi.org/10.1371/journal.pgph.0004223.t001

Prevalence of sexual violence experiences

The prevalence of SV experiences among adolescents is presented in Table 2. The results indicated that a majority (69.4%) of adolescents experienced some form of SV since the age of 12. The spectrum of experiences ranged from non-contact abuse (63.2%) to passive contact abuse (41.9%), active contact abuse (28.7%), forced intercourse (21.3%), and transactional sex (14.9%). Substantial differences emerged between in-school and out-of-school adolescents, as well as between boys and girls. Out-of-school adolescents consistently reported higher instances of SV across all categories, while boys, irrespective of school attendance, reported significantly more SV experiences.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Prevalences of sexual violence.

https://doi.org/10.1371/journal.pgph.0004223.t002

Perpetration of sexual violence

When analysing perpetrators of sexual violence (SV) by gender, different patterns emerged. Peers were significant perpetrators for both males and females, but with notable differences in the specific types of peers involved.

Male survivors.

For male survivors, female peers emerged as the main perpetrators in most forms of SV (Table 3). Female peers at school (24.4%) and other female peers (23.2%) accounted for the highest percentages of perpetrators for all forms of sexual abuse. This pattern is particularly pronounced for active contact abuse, where female peers at school (37.3%) and other female peers (29.5%) are the dominant perpetrators. Male peers also play a significant role, especially in non-contact abuse (27.1% for male peers at school and 22.2% for other male peers). However, their involvement is generally lower than that of female peers. Adults, both male and female, are less frequently reported as perpetrators by male victims, with percentages typically ranging from 5% to 10% across different forms of abuse.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Perpetrators of sexual violence disaggregated by gender (male and female).

https://doi.org/10.1371/journal.pgph.0004223.t003

Correlates of sexual violence

The associations between adolescents’ characteristics and SV experiences are explored in Table 4. The baseline model, including gender, age, location, intimate partnership, and SV before the age of 12, revealed significant associations. Being male, older, and involved in a romantic relationship were all linked to a higher SV risk. Remarkably, the strongest predictor of SV was experiencing SV before the age of 12 (OR 4.622), followed by being in a romantic relationship (OR 2.731). Boys faced a higher risk than girls (OR 2.033). Our analysis of additional potential correlates and their significance demonstrated that the baseline model accounted for the higher prevalence of SV among out-of-school adolescents (column 2). Household wealth (columns 3&4) emerged as a significant factor, with each additional household asset reducing the SV risk (OR 0.902). Notably, adolescents who reported receiving emotional support from both parents (column 5) experienced a lower risk of SV (OR 0.413).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Correlates of sexual violence.

https://doi.org/10.1371/journal.pgph.0004223.t004

Regarding gender interactions in the regression model (S8 Table. Urban/rural gender interactions - Appendix), the urban/rural effect becomes significant and protective when the interaction is taken into account (OR 0.579). This suggests that, overall, living in a rural location is associated with a lower likelihood of experiencing sexual violence. The significant interaction (OR 2.529) indicates that this location effect differs by gender, with the protective effect of an urban location (for males) being reversed (for females) due to the interaction effect.

Perpetration

Our gender-disaggregated analysis revealed different patterns of perpetration for male and female survivors of sexual violence. While peers are important perpetrators for both genders, male survivors are more likely to report female peers as perpetrators, particularly in cases of active contact abuse. In contrast, female survivors predominantly reported male adults and peers as perpetrators, especially in cases of forced intercourse. These findings highlight the need for gender-specific approaches in sexual violence prevention and intervention strategies, with a particular focus on peer-to-peer dynamics and adult-youth interactions. Often neglected, peer victimization have proven acute and lifelong consequences on mental health [43], with an increased risk of depression [44], and anxiety disorders [45]. For our sample, Doerr et al.[2] showed that adolescents who experienced the highest levels of SV also exhibited strongest negative impact on their mental and physical health. The high rates of peer perpetration (77.1%) observed in our study have significant implications within the context of Nigerian social structures and adolescent development theories. This finding is consistent with the social learning theory, which theorises that adolescents may learn and internalise violent behaviours through observation and imitation of peers [46]. In the Nigerian context, where hierarchical social structures and gender inequalities persist, peer-perpetrated SV may reflect broader societal norms and power dynamics. Moreover, the social ecological model of adolescent development shows the pivotal role of peer influences during this formative period [47]. The high prevalence of peer perpetration indicates a need for interventions that target peer groups and social norms, as opposed to focusing exclusively on individual-level factors or adult perpetrators. Such an approach could include the implementation of peer education programmes, the provision of bystander intervention training, and the introduction of initiatives that promote healthy relationships and gender equality among adolescents [48].

Gender disparity

A prominent observation from this study was the significant gender disparity in SV experiences, with boys consistently reporting higher rates of SV across all categories compared to girls. While the VACS study [21] presented prevalence rates of SV for boys, it is noteworthy that the academic literature has generally paid limited attention to boys’ experiences in Nigeria, as evidenced by Nguyen et al.’s study [29]. The higher reported rates of SV among boys in our study prompt us to consider the need for a more detailed understanding of this issue, which has predominantly focused on female survivors. Many factors might influence this discrepancy, such as, cultural norms and reporting biases. In many Nigerian contexts, traditional gender roles may discourage boys from reporting SV due to fear of stigma or being perceived as weak, which could potentially lead to under-reporting in other studies [49,50]. Additionally, it is possible that societal norms, which emphasise male dominance and aggression, may contribute to the normalisation of SV against boys, which in turn may make it less likely to be recognised or reported as an abuse [51]. This emphasises the pressing need for gender-specific interventions that address these cultural barriers and promote a supportive environment for all survivors, regardless of gender [50]. Thus, it would be beneficial for policies to place an emphasis on increasing awareness about male victimisation and providing targeted support services that are sensitive to the unique challenges faced by male survivors.

Correlates of risk and protective factors

The correlates of SV experiences showed key risk and protective factors. Being male, older, and involved in a romantic relationship were all associated with a higher risk of SV, aligning with previous research on vulnerability factors [31]. Remarkably, the strongest predictor was experiencing SV before the age of 12, suggesting a cyclical pattern of victimization and the need for an early intervention and support services.

On the other side, household wealth emerged as a significant factor, with higher socioeconomic status acting as a protective factor against SV. This finding underscores the role of poverty and economic vulnerability in exacerbating the risk of transactional abuse, particularly in resource-limited communities [31,52]. Previous studies disagreed on the importance of wealth. Yahaya et al. [53] found no correlation between socioeconomic status and childhood sexual abuse, while Miller and colleagues [54] suggested that children aged 13-17 years in households with high economic status are more likely to experience SV. Our data provided a more detailed insight. We found that although children from poorer households were more likely to experience SV, there was a threshold effect. The risk of SV is only significantly lowered for those from relatively wealthy households (with seven or more household assets, which applies to 28.7% of the adolescents interviewed).

Additionally, emotional support from both parents was associated with a lower risk of SV, highlighting the importance of family support systems and positive parenting practices in mitigating adolescents’ vulnerability. Therefore, in line with previous research [55], our results emphasize the impact of both parents emotional support correlated with lower probabilities of experiencing SV, transcending the significance of wealth in preventing SV. It may be beneficial to consider programmes that enhance parental involvement and communication, with the aim of fostering a supportive home environment where adolescents feel safe to discuss their experiences and seek help [56]. Furthermore, encouraging positive parenting practices and providing resources for parents to recognise signs of SV, could further enhance these protective effects [57].

This study also revealed that out-of-school adolescents consistently reported higher frequencies of SV across all categories compared to their in-school counterparts, and reported the highest rates of (poly)victimization. This included an elevated risk for non-contact abuse, contact abuse, and forced intercourse when compared to both in-school girls and boys. Few studies presented similar findings [58,59]. One study involving South African students revealed a higher prevalence of sexual violence among boys, with rates of 60% for boys compared to 53.7% for girls [59]. Targeted outreach and non-formal education programmes can be developed to reach out-of-school adolescents, especially in underserved communities. These programmes could offer flexible learning options, vocational training and life skills education tailored to their needs and circumstances. For example, the Association for Reproductive and Family Health (ARFH) in Nigeria has conducted outreach activities targeting out-of-school youth, including apprentices and artisans, providing training to empower them and help them make informed decisions about their sexual health [60]. In addition, accelerated learning programmes, such as those supported by Education Cannot Wait in northeast Nigeria, can help out-of-school adolescents catch up on missed education and potentially reintegrate into formal schooling [61]. Community-based interventions such as the WARIF Educational School Programme (WESP) could be adapted to reach out-of-school youth, focusing on building capacity to respond appropriately to and prevent sexual and gender-based violence [62]. In addition, integrating sexual and reproductive health education into these programmes, as demonstrated by projects such as the Life Planning Education Project, can address critical health needs of this vulnerable population [60]. These initiatives are a collaboration between government agencies, NGOs and community organisations, targeting the development of comprehensive strategies that address the multiple challenges faced by out-of-school adolescents, including poverty, early marriage and lack of access to health services.

Urban-rural dynamics and gender interactions

Our analysis revealed complex interactions between gender and urban-rural setting in relation to sexual violence risk. When taking gender interactions into account, living in an urban area emerged as a protective factor against sexual violence (OR 0.579). However, this protective effect varied significantly by gender (interaction OR 2.529). The results suggest that the urban environment may be more protective for boys, while the protective effect may be reduced or possibly reversed for girls. This nuanced relationship highlights the importance of considering gender-specific experiences in different geographical contexts. Urban areas may offer more protection to boys through better access to education, support services and potentially more progressive social norms. Conversely, girls may face unique risks in urban settings, such as increased exposure to diverse social interactions, potential anonymity in larger populations, or specific forms of gender-based violence that are more prevalent in urban settings. These findings are consistent with previous research that has highlighted the complex interplay between urbanisation, gender and violence [63,64].

It is important to keep in mind that any correlates of intra- and interpersonal factors with violence victimisation/perpetration and derived patterns of violence need to be seen in context. In order to better understand the relevance of such factors for individual children and adolescents and their environment, we need more contextualising research approaches such as qualitative in-depth interviews and focus groups discussions.

Patterns of violence

The analysis of different forms of SV revealed diverse patterns. For instance, boys faced higher risks of passive contact, active contact, and non-contact SV. The data indicated that older adolescents reported a greater number of experiences of both active and non-contact abuse. It is important to note, however, that this observation may be attributed to the cumulative nature of our measures (older, longer the exposure) rather than indicating increased vulnerability with age. The presence of a romantic partner amplified the risk across all SV forms, underscoring the potential for intimate partner violence. There was a substantial prevalence of non-contact abuse across the sampled groups: in-school girls (54%), in-school boys (72%), out-of-school girls (80%), and out-of-school boys (92%). This observation aligns with previous research by Rostad et al. [65] and Peter and Valkenburg [66], indicating a prominent link between pornography exposure and SV perpetration and victimization. This is particularly concerning in the digital age, where adolescents are increasingly exposed to online risks.

Strengths and limitations

The study’s strengths include its regional representativeness, inclusion of both in- and out-of-school adolescents, and the use of validated measures and rigorous data collection procedures. In addition, various forms of SV were discerned through more detailed categorization: non-contact SV and contact SV, further disaggregated into passive and active contact abuse, forced intercourse, and transactional sex. However, it is important to acknowledge some limitations. First, the cross-sectional nature of the study excludes the establishment of causal relationships between the identified risk factors and SV experiences. Second, the out-of-school subsample was relatively small, and it was collected in very challenging circumstances. Due to safety concerns for the enumerators, data from the out-of-school adolescents was collected using pencil and paper rather than tablets, therefore these questionnaires were not self-reported. Asking face-to-face may have led to different disclosure rates, although existing evidence on this matter is inconclusive [67,68]. Third, requesting parental consent could potentially have led to under-reporting of SV situations carried out by parents or other family members. Finally, another limitation of our study is the lack of formal power calculations to determine sample size, which we strongly recommend for future research. Instead, our sample size was based on practical considerations, resource constraints and logistical factors, while aiming to reflect the population distribution of in-school and out-of-school youth in the study area. However, the present findings can guide future studies in their power calculations.