Learners’ perspectives on school safety in Johannesburg

In this article we report on a South African study conducted with 1,779 learners aged 11 to 18 years from 8 schools in urban areas in and around Johannesburg. These learners’ perspectives on school safety confirm that South African learners experience their schools as unsafe. Their primary concerns related to coercion and violence against learners by peers or teachers on school grounds, although they regarded the physical school environment and domestic and community conditions as impacting school safety. To address safety concerns, learners desired the punishment of offenders and greater involvement and accountability of adults. We recommend an urgent whole-school intervention using a critical gender lens.


Introduction
Schools are a "site of violence, [a] producer of violence, and [a] starting point for ending violence" (Morrell, 2002:39).
Many South African schools are places of physical and emotional vulnerability and danger (Burton & Leoschut, 2013;Ncontsa & Shumba, 2013;Van Jaarsveld, 2008;Zuck, Mathews, De Koker, Mtshizana & Mason-Jones, 2017). Unsafe schools disrupt both the education and broader development of those affected. In contrast, learners thrive in safe, caring schools where both the fear of, and actual threats to, safety are absent (Burton & Leoschut, 2013;Darling-Hammond, Flook, Cook-Harvey, Barron & Osher, 2020). Using a gendered lens, in this article we report on a South African study in which we examined the perspectives of learners regarding their biggest worries at school. With the study we aimed to address the paucity of learner voices in school safety research, and to inform the services and advocacy efforts of the Teddy Bear Foundation (hereafter referred to as the Foundation), a non-governmental organisation treating and preventing child abuse. We offer recommendations framed by the participating learners' voices, strengthening the evidence-based informing school-safety interventions (Burton, 2008) and supporting the advancement of caring and safe school environments. We conclude that a whole-school, multi-dimensional approach (Kreifels & Warton, n.d.;Parkes, Heslop, Johnson Ross, Westerveld & Unterhalter, 2016;SaferSpaces, 2022) informed by a critical gender lens is needed to shift schools towards greater equity, peace and safety.

Conceptual Framework: Framing School Violence as Gendered
In this article we pay particular attention to a gendered experience of school violence. Alongside international colleagues, many South African researchers resist understanding school violence in individualistic or psychological terms, because this misses the social and structural roots of the phenomenon (Bhana, De Lange & Mitchell, 2009;Bhana & Pillay, 2011;Morrell, 2002). Noting that the school system has been historically designed to control and discipline populations (Gopal & Collings, 2017), gender is one systemic aspect to be considered. Indeed, schools mirror the gendered power dynamics of broader society (Morrell, 2002;Taole, 2016). Institutions, including schools, can challenge gender violence and enhance safety by consciously creating a space of equity and care, and by attending to the links between violence, identities, socio-cultural norms and intersecting structural inequalities (Parkes et al, 2016).

Literature Review: School Violence in South Africa
Safety at school is a dominant concern in South African education. Indeed, around 22% of learners report experiences of violence in the school context (Burton, 2008;Burton & Leoschut, 2013;George, Finberg & Thonden. 2001), some having been victimised multiple times (Leoschut & Kafaar, 2017). Although South Africa is not unique in worrying about school safetyfor example, much of the literature discusses the United Stateswe rely primarily on local literature. This allows us to highlight the construction of South African school safety/violence in its unique context.

Dimensions of violence
School violence is a complex and multidimensional phenomenon. Interpersonal violence, community violence and the lack of adequate physical infrastructure intersect in creating safety issues at schools. Interpersonal violence includes sexual assault, sexual harassment, bullying and corporal punishment (Lester, Lawrence & Ward, 2017). Boys are more likely to be victims of sexual assault at primary school (Burton, 2008), but girls are especially vulnerable to sexual harassment, rape, verbal degrading and assault in secondary schools by male peers and teachers (George et al., 2001;Jewkes, Levin, Mbananga & Bradshaw, 2002;Mayeza & Bhana, 2017;Prinsloo, 2006). Such behaviour is legitimised through strong beliefs in male power. The violent performance of hegemonic masculinities (Morrell, 2002) is reflected in male teachers' and learners' assumptions that men can and should control womentheir bodies and societal placeand should control males who do not fit narrow, dominant ideas of masculinity (Bhana et al., 2009;Carter, 2002;Parkes, 2015). Such behaviour is further reinforced by a general disregard for girls' rights (Bhana, 2012;Prinsloo, 2006). Learners who identify as lesbian, gay, bisexual, transgender, and questioning (LGBTQ) face increased violence (Daniels, Struthers, Maleke, Catabay, Lane, McIntyre & Coates, 2019). Bullying is also frequently marked by dominant forms of masculinity, while violence between girls follows patterns of social power and heterosexual competition (Bhana & Pillay, 2011). Even while some teachers speak of the need to teach respect and peace (Mayeza & Bhana, 2017), authoritarian teaching methods and fear of losing control result in consistent accounts of moderate to severe forms of physical punishment and humiliation (Breen, Daniels & Tomlinson, 2015;Burton, 2008;Mayisela, 2018;Seedat, Van Niekerk, Jewkes, Suffla & Ratele, 2009). A gendered lens is congruent also with the perception that corporal punishment best controls disobedience, bad behaviour, and a lack of academic progress (Mayeza & Bhana, 2017;Tafa, 2002).
Teachers and administrators have themselves been verbally abused or assaultedsometimes by learners (Burton, 2008;Ncontsa & Shumba, 2013). This again follows a gendered pattern of aggression male learners targeting teachers to assert their dominance (Bhana & Pillay, 2011;Carter, 2002). Some teachers may respond to such helplessness and frustration with increased aggression towards the learners (Shields, Nadasen & Hanneke, 2015).
Beyond interpersonal violence, the high rates of community violence, a lack of community safety, and inequityagain experienced disproportionately by South African femalesextend into the school experience (Breen et al., 2015;Burton & Leoschut, 2013;Mayeza & Bhana, 2017;Seedat et al., 2009). This occurs, for example, when gang differences, the availability of weapons, or community conflict drive learner violence (Ncontsa & Shumba, 2013). Additionally, fractures in family life (such as those related to domestic violence or substance use, or absence of family members due to incarceration or migration) create further vulnerability for learners (Burton & Leoschut, 2013). Learners are also made unsafe by prevalent vandalism and robbery (Ncontsa & Shumba, 2013).
Another issue contributing to a lack of safety and risk, particularly for girls, is poor physical infrastructure such as inadequate toilet facilities (often the site of sexual assault (Draga, 2017)), a lack of classroom furniture and books, and overgrown or dangerous playgrounds (Xaba, 2014).

Multifaceted effects
Exposure to such violence has multifaceted effects impacting the physical, psychological, social, behavioural, spiritual, and cognitive development of learners in the short and long term (Jewkes et al., 2002;Meinck, Cluver, Boyes & Loening-Voysey, 2016). The list of direct and indirect impacts includes concentration struggles, developing negative identities, adopting violence to address conflict, incurring physical injury, feeling guilty or ashamed, displaying mental health and behavioural challenges, engaging less in the classroom, attending school less often, and increasing the use of substances (Breen et al., 2015;Burton, 2008;George et al., 2001;Ncontsa & Shumba, 2013;Ward, Artz, Leoschut, Kassanjee & Burton, 2018). In turn, learners may have poorer grades; and may be repeatedly exposed to the perpetrators, risking secondary victimisation (Meinck, Cluver, Boyes & Mhlongo, 2015). Teachers can feel that their ability to teach effectively is undermined and constrained, resulting in them spending much time attempting to manage conflicts (Ncontsa & Shumba, 2013). Thus, fear and the lack of a cohesive community can spread through a school setting even if not everyone has been directly affected (Burton & Leoschut, 2013). Noting that girls are more likely to be subjected to violence, they may also be more vulnerable to these impacts on their learning.
While outlining the safety challenges in schools, the literature suggests that successfully transforming a school into a safe place has dramatic short, medium-and long-term positive effects on individual, social and community outcomes. Such schools cement cohesion among learners and between learners and teachers, create a supportive learning environment, advance engagement in the curriculum, enhance confidence, develop solidarity, promote curiosity, teach constructive conflict resolution, and promote fun in the learning environment (Battistich, Solomon, Watson & Schaps, 1997;Burton & Leoschut, 2013;Sulley, 2007). Reducing risk factors at school level can build individual and social resilience and have a "synergistic impact on other spheres of protective factors" (Burton & Leoschut, 2013:5), such as in the family or community.
School violence has been addressed through specific, albeit limited, initiatives (Palm, 2019). The Department of Basic Education has developed a whole-school National School Safety Framework in 2015 to address school safety (SaferSpaces, 2022). Provincial programmes include the Sexual violence against young girls in schools in South Africa project and the STOP, WALK, TALK antibullying programme (Kreifels & Warton, n.d.). Some safety programmes pair schools with police stations. Despite some promising outcomes, school safety for all learners, female learners in particular, remains a major concern.
Methodology As part of the Global Action Day for Basic Human Rights to Health in 2014, the Foundation, which has a long history with local schools, embarked on a study to access learners' views regarding school safety. The Foundation focussed on learners' voices, because learner input is often neglected in school safety research (Burton, 2008). Based on these perspectives, the Foundation planned to use such evidence-based information to assess the relevance of the interventions they offered to schools and towards advocacy.
Using convenience sampling and taking into consideration the agency's capacity, the Foundation selected the first eight schools in the metropolitan region of urban Johannesburg that responded with interest to a telephonic invitation. These schools were predominantly poor and primarily populated by black learners, most of whom were disadvantaged.
At each participating school, the Foundation's social auxiliary workers presented a talk to all the schools' learners on child abuse. They also invited learners between 11 and 18 years to anonymously and voluntarily answer two questions on a handout: "What is your biggest worry at school?" and "What would you like to see done about this?" To facilitate anonymity and participation, learner responses were deposited into unmonitored foundation collection bins available at each school for 2 days. Learners were verbally informed that their responses would be used to allow the Foundation to better understand needed changes to increase school safety and to advocate on the learners' collective behalf. These written responses provided a qualitative dimension that presented learners' experiences in their own words.
The researchers adopted both qualitative and quantitative approaches to analysing the data. Counting the number of coded responses offered an indication of trends. Written responses were analysed through a process of thematic content analysis, starting with identifying common threads, sorting these into categories, and then into themes. These themes revealed textured insights into the learner responses.
The study was limited in various ways. Anonymous, written responses could not be further explored or corroborated through member checking. While aiming for simplicity and thus not asking the learners to provide any demographic information, we were forced to make broad inferences regarding, for example, gender. The research was also problem-oriented, and deficit based, identifying solely learners' worries and not identifying what might be working for learners.

Findings
A total of 1,779 learners responded to the questions. This represented 27% of the total school population surveyed and provided a reliable indication of the learners' worries and recommendations. The participation range varied across individual schools from 11% to 42%, seemingly dependent on the availability of the learners at the time that the talk was held. Learners sometimes offered more than one worry, all of which were captured.
What is Your Biggest Worry at School?
The dominant theme that emerged in response to "What is your biggest worry at school?", related to concerns about abusive behaviour at school (teacher and learner behaviour), followed by unsafe behaviour by other learners in the school environment.

Abusive behaviour by teachers and learners
Many learners raised abusive behaviour as a significant worry at school. The quotes tell their own story: A quarter of the respondents suggested that they had experienced or witnessed bullying by peers such as their peers being beaten/beating them, removing food, stealing stationary from and insulting other learners. South African studies have reported rates of up to 80% of bullying both at, or on the way to, school (Burton & Mutongwizo, 2009;Mahabeer, 2020;Rubbi Nunan & Ntombela, 2019). Where the gender of the bully or the victim was identified in the Foundation's study, perpetrators were typically boys, and the victims most often girls, corroborating the gendered nature of school bullying (Bhana & Pillay, 2011;Carter, 2002;Morrell, 2002). Cyberbullying was an issue at one school only, although this merits further investigation noting that there is high use of mobile phones in all South African schools (Burton & Mutongwizo, 2009, Kritzinger, 2017. Learners were concerned about teachers being physically and emotionally abusive through corporal punishment, denigrating language and extorting money from the learners. Where the gender of an abusive teacher was identified (only once), it was a female teacher. Some learners reported worries related to school rules, for example, what hairstyles or dress could be worn to school. Such issues have a safety dimension in that they may be related to the exercise of discipline. Similarly, reports by some learners of the fear of failing and not receiving the expected support or interest from teachers referred to a negative learner-teacher relationship.
A very small number of learners reported sexual abuse. Sexual abuse is believed to be extensive, although underreporting is common (Burton & Leoschut, 2013;Jamieson, Sambu & Mathews, 2017), and seemed to prevail even in the Foundation study where anonymity was ensured. No learners identified bullying, or abuse, related to lesbian, gay, bisexual, transgender, queer (or sometimes questioning), and other (LGBTQ+) statusthis silence perhaps being related to the societal taboo nature of this topic (Daniels et al., 2019).

Unsafe learner behaviour
Learners were distressed by the disruptive and unsafe behaviour of fellow learners. Concerned about being exposed to substances at school, one learner observed: "My biggest worry at school are [sic] the children who are smoking all sorts of drug substances and the problem is they are smoking them anywhere they like and it's affecting us." This supports other observations that substance use is prevalent among South African adolescents, (Reddy, Resnicow, Omardien & Kambaran, 2007). It is a concern associated with exposure to violence, early sexual activity, and depressive symptoms (Magidson, Dietrich, Otwombe, Sikkema, Katz & Gray, 2017). Other learner behaviour that resulted in learners feeling unsafe included disrespectful behaviour directed at both teachers and their peers, stealing, gambling, teenage pregnancy, and sexual activity on the school premises. Gendered dimensions emerged in the comments regarding pregnant learners that conveyed prejudices against such female learners, again reflecting dominant biases (Matlala, Nolte & Temane, 2014). It is assumed that sexual activity reflected some exploitation of girls (Bhana, 2012;Morrell, 2002).

Further concerns
My biggest worry at school is that some people get sick because of the dirty toilets. Safetywe are not safe we have only one security.
The literature highlights that poor school infrastructure intersects with interpersonal and community violence as a dimension of the lack of safety at schools. This issue was also reflected in the study. Some participants cited inadequate school facilities (such as dirty and unhygienic toilets, litter and dangerous rubbish, unmaintained grounds, and insufficient desks and seats) as compromising their health, placing them at risk of bullying, or preventing participation in recreation. It would seem that unsafe toilet facilities placed girls and non-binary gender learners at greater risk than their male peers (Daniels et al., 2019). Learners at a city school felt that there were too few security guards protecting the school, highlighting their fears of neighbourhood violence spilling onto the school grounds. For a small number of learners, an unsafe home environment impacted their sense of safety at school and thus was a worry. Issues ranged from physical, sexual and emotional abuse at home, to inadequate provision because of unemployment. One learner asserted: "I'm not able to concentrate in class and I always feel like sleeping … this happens because when I am at home I can't sleep because I am afraid of my big brother." Some learners were worried that the school nutrition schemes were not providing nutritious, or enough, food. While this might not be regarded overtly as a safety issue, sufficient and nutritious food enhances concentration and engagement, thus creating more robust learners and reinforcing the message of a caring, safe school.

No concerns
Four per cent of the responses indicated no concerns in the school environment. Considering the multitude of concerns raised, one wonders what allowed these learners to feel that their needs for learning and safety were being adequately addressed.

What Do You Think Should be Done?
The learners offered several suggestions to enhance school safety, such as punishing offenders and greater involvement of adults.
I would like the teacher to do something to bullies [to make them stop]. I would like those children to be suspended from the school forever. Please tell this teacher to stop beating us and swearing at us and give her a warning. The teachers must be fired or get suspension or expelled or get a written warning. I would like to see the police or their mothers to talk to these boys, [because] we are sick and tired of them. So please help us and please stop them what they are doing.
The above quotes suggest that learners commonly felt there should be punishment, implying a worldview that considers punitive measures to be the best response to problems. Such a perspective is consistent with a punitive societal view and an authoritarian and punitive school culture (Harber, 2002).
Additionally, the responses indicated frustration regarding a lack of accountability by the responsible adults who should take control and create safety. Learners felt that the abuse should be stopped by the police, school authorities should deal with errant teachers, parents should be more involved, the Foundation should provide further counselling and support, and teachers should offer learners greater attention. Learners also recommended increased access to food, new facilities and access to resources. Taken together, these responses offered a relatively nuanced platform for action.

Discussion
Three potential conclusions emerged from the findings. Firstly, the learners presented a concerning picture of their everyday school experiences, using their own words to conclude that schools were not safe places for them. Having this asserted by learners themselves is notable as previous research on violence in South African schools has mostly overlooked learners' voices. Interestingly, however, while the findings corroborate outcomes regarding learner needs from other South African research, they seem to underidentify violence in their schools, especially in relation to sexual assault (Bhana, 2012;Jewkes et al, 2002;Meinck, Cluver, Loening-Voysey, Bray, Doubt, Casale & Sherr, 2017). This underrepresentation may be related to the study methodology and questions being framed in the language of worries, as well as the taboos related to certain types of violence. However, the findings do not weaken the assertion that schools are unsafe places for learners.
Secondly, learners' responses reinforced notions of the complexity of school safety. Learners implied that they functioned within an especially insecure space for teaching and learning. Although learners prioritised issues relating to physical integrity and social and psychological safety on school premises, they also noted that the physical infrastructure as well community and domestic conditions impacted their experience of school safety. The toxic mix of an unsafe school environment coupled with the socio-economic disadvantages, community and familial violence experienced by most of these learners resulted in a strong likelihood of a lack of educational progress, school dropout, and lower grades (Herrero Romero, Hall & Cluver, 2019;Pieterse, 2015;Sherr, Hensels, Skeen, Tomlinson, Roberts & Macedo, 2016). Such an unfavourable environment may undermine educational inputs and policies put in place by the state. Learners' voices thus suggest that while the emphasis needs to be on addressing interpersonal violence, the school environment and ways of mediating the spill-over of risk and insecurity from the community and home additionally warrant attention. Learners also offered explicit recommendations. Although these initially appeared to be focused on punishment and police involvement, thus echoing the larger societal discourse, a more careful reading of learners' comments suggests a complex, multi-layered palette of desired interventions. Learners looked to the adults in the school system to take responsibility and promote accountability and identified roles for each of these stakeholders. In asking for greater teacher involvement and attention, and hoping for further support from the Foundation, learners seemed to regard care and safety as mutually reinforcing. The learner's recommendations thus underlined a whole-school response.
Thirdly, a gendered lens to understanding school safety issues was indirectly reinforced. Although we did not ask participants to identify their own gender nor that of those referred to in their responses, a gendered analytical lens suggests that a learner's gender determines how safety is enacted in the school context. In the instances where gender was identified, perpetrators were mostly males. Sexual harassment and abuse were largely directed at females. Learners appeared to adopt binary notions of gendered safety, perhaps as inferred earlier due to the societal silence regarding a range of gender identities. Even with the methodological limitations, the learners' voices thus echoed the findings elsewhere of the gendered nature of the violence in local schools (Bhana, 2012;Morrell, 2002). Gendered violence in schools perpetuates the normalisation of exploitive, domineering, and violent behaviour. Where gendered power relations are not explicitly challenged or addressed, learners have little opportunity or skill to shift these damaging social processes, now or in their future (Carter, 2002). This process may hinder the ability of learners to form healthy and trusting relationships with peers and adults, encouraging aggressive or violent behaviour, and reducing emotional resilience (Burton & Leoschut, 2013;Herrero Romero et al., 2019).
Increased surveillance, the securitisation of schools (Xaba, 2014) and severe punishment is an attractive solution in the face of apparent rampant ill-discipline, a perilous environment, and emergent fear. However, such repressive measures have been shown locally and internationally to be counterproductive, while in contrast, facilitating inclusion and utilising restorative justice approaches contribute to both school safety and academic achievement (Gevers & Flisher, 2012;Kupchik, 2016). Following learners' insights, a whole-school approach involving all stakeholders (Burton, 2008) is required for school safety. This includes engaging caregivers and community stakeholders, because this not only increases school safety, but expands safety into learners' homes and their communities (Gevers & Flisher, 2012;Prew, 2009). School safety interventions need to be integrated, cohesive, multi-levelled, comprehensive and sustainable (Burton, 2008;Prinsloo, 2005), and should promote accountability, good leadership, governance and support for teachers (Burton, 2008;Prinsloo, 2005;Shields et al., 2015). Moreover, a whole, caring-school approach recognises the reality of poly-victimisation and offers multilevelled, culturally appropriate interventions that address prevention, enhance protective factors and link to social determinants of individual and social health (Lester et al., 2017;Maringe & Moletsane, 2015;Meinck et al., 2015).
Doing gender work with teachers to subvert and interrupt "the complex edifice of polarized gender hierarchies at the structural and cultural level" (Bhana et al, 2009:59) is critical in order to change harmful discourses and behaviour perpetuated by both learners and teachers. Noting that even in an environment of reluctant disclosure learners are most likely to confide in caregivers and teachers (Meinck et al., 2017), it is important to enable these adults to be responsive to learners' concerns. Reporting mechanisms must be simple and easily accessible (Burton & Leoschut, 2013;Ncontsa & Shumba, 2013). The observation in the literature that boys may be more vulnerable to sexual abuse in primary school alerts policy makers to the need for adopting differentiated approaches to gender-based violence. Also, gender must be understood in its complexity, avoiding binary approaches.

Conclusion
Zeroing in on learners' voices, we emphasised the vulnerability of learners in the school context. We underlined the value of a whole-school response that addresses the complexity of school safety and engages stakeholders (Lester et al., 2017). Furthermore, we highlighted the importance of recognising the role of gender in creating safe school environments because the most wellmeaning programmes will be unlikely to succeed if gendered expectations of social power and privilege are not effectively challenged. A critically informed whole-school, gendered approach can powerfully interrupt the cycle of violence and create environments of care, promoting belonging, safety, trust and cohesion.
The Foundation acted on this study by presenting the results to the provincial Department of Education, which subsequently agreed to follow up specific concerns in certain schools. The Foundation has also developed comprehensive recommendations for designing and implementing a whole-school response to school safety; delivered further services to schools including talks to learners regarding rights and responsibilities, safe and unsafe touching, information regarding support, and established girls' leadership clubs. The Foundation further encourages such initiatives.