Risk Factors of Sexual Abuse among School going Children in Primary Schools in Kisumu County, Kenya

Child sexual abuse is a public health problem with negative physical, mental, behavioral and learning outcomes to survivors. Recent Kenya Demographic Health Survey showed that 22% of women and 13% of men in western Kenya have experienced sexual abuse compared to 14% women and 6% men nationally. Its burden, predisposing factors and, sources of or reasons for differential exposure for boys are yet unclear. Whereas schools have a major inuence on a child’s development and behavior, they have been identied as important settings where child sexual abuse might occur. However, how sexual abuse is patterned by type and location of school remains an open question. The present study investigated risk factors of sexual abuse among school-going children in primary schools in Kisumu County. Using a cross-sectional study design and guided by the ecological model, the study determined association between type of school and sexual abuse; association between location of school and sexual abuse, and spaces that predispose school-going children to sexual abuse in primary schools. A sample of 398 pupils aged between 10 – 16 years randomly identied participants from 37 pre-selected primary schools in Kisumu Central and Kisumu West sub-Counties was used. The schools had been stratied by location (rural and urban) and type (public or private, boarding or day). Parental permission was sought for respective pupils to participate in the study. Data from pupils was collected using self-administered questionnaires and focus group discussions, and key informant interviews from guidance and counseling teachers. The mean age of participants was 13 years. More pupils attended day schools than boarding schools (88% versus 12%; p= 0.017). Forced sex (29%) was the most common type of sexual abuse. More girls than boys had ever experienced any type of sexual abuse 29% versus 20%; p=0.038. Whereas there was no statistical difference in pupils’ experience of sexual abuse by type and location of school x2= 2.044; p>0.05 and x2= 0.823 p>0.05, pupils thought that attending public day schools is associated with higher risk for sexual abuse mainly while walking to and from home. Pupils cited bushes inside school (52%), latrines/toilets (51%) and bushes outside school (50%) as spaces where one is more likely to experience sexual abuse. Being a female pupil was the most signicant risk factor for sexual abuse in the primary schools surveyed. This indicates that gendered predisposition plays an important role in occurrence of sexual abuse in primary schools. Though not statistically signicant in this study, it is plausible that risky spaces and type of schools might modify female gender predisposition in subtler ways as suggested from the pupils’ perspectives. There is need for a gendered focus in a complex primary schooling environment


Introduction
Childhood maltreatment is a serious global public health problem with long-lasting negative consequences on the mental, physical and social wellbeing of survivors (Kumar,Singh & Kar, 2017). Globally it is estimated that up to 1 billion children aged 2-17 years have experienced physical, sexual or emotional violence or neglect (WHO, 2018). Worldwide, the most recent surveys indicate that 9 million girls aged 15 to 19 were forced into sexual intercourse or other sexual acts within the past year. Moreover, girls report lifetime rates three times higher than boys in high income countries (UNICEF,207). Although boys are also at risk, global estimates on the burden of sexual abuse on boys are unavailable (Zalewski, 2018;UNICEF, 2017).
Despite the high burden of sexual abuse and its adverse effects, child sexual abuse is one of the most underreported crimes (Mathews, Brom eld, Walsh, Cheng & Norman, 2017). Research has attributed underreporting of sexual abuse to shame, guilt and stigma of sexual victimization (Johnson, 2012;Sable, Danis Fran, Denise & Sarah, 2006), normalization of inappropriate or unwanted sexual behavior, and perception that sexual violence does not warrant reporting (Linda 2018). In addition, differences in de nitions of sexual abuse across cultures or countries are thought to negatively impact its reporting (UNICEF, 2012;Renold 2002). In Kenya the frequent cases being reported to the police, rescue centers and in the media paint a bleak picture of the situation. For example, 2010 national police statistics revealed an average of over 200 cases of de lement reported every month in the country (Kenya Police Service, 2010).
Schools have a major in uence on a child's development and behavior, yet they have been identi ed as one of the settings where child sexual abuse occurs. For instance in private schools, pro t-making has been the motivation at the expense of a child's welfare (Royal Commission, 2017). In Kenya, after the government introduced free primary education in 2002, public schools faced congestion with older mature kids being reenrolled back to school. More so, public schools are regulated by the Teachers Service Commission while private schools are not. However, whether such differences underlie differences in the burden of child sexual across type of schools (public vs. private) remains unknown. Furthermore, with regards to type of school, boarding school pupils provide opportunities for increased interaction with the school staff and other pupils and not parents or guardians. In contrast, learners in day schools potentially experience abuse while transiting to and from school. However, whether such differences between day and boarding schools manifest in differences in the burden of sexual abuse among primary school children is not known.
In addition to differences in terms of risk factors between day and boarding schools, children attending schools in urban areas may face different risks to sexual abuse compared to those attending schools in rural schools. Studies conducted in Zimbabwe and Ethiopia showed that pupils in rural schools were more at risk of facing child sexual abuse than those in urban schools and these differences were attributed to how knowledgeable of child sexual abuse such that children in rural schools were found to be less knowledgeable on the subject (Shumba, Gwirayi, Maphosa, Gudyanga & Makura, 2015, Mat, 2016. Children attending rural schools are exposed to sexual abuse as they travel long stretches of bushy and forested paths early in the mornings and late in the evenings (Kenya Catholic Secretariat of Religious Education, 2015). Children attending schools in urban areas also face their own set of risks. For example, issues of drugs and alcohol abuse and peer in uence, which are known risk factors to sexual abuse are more prevalent in urban areas than in rural areas. In addition, abandoned or neglected buildings and dumping sites that are common in urban areas may expose school-going pupils to sexual abuse. However, empirical evidence is not available on whether there indeed is an association between burden of child sexual abuse and location of school de ned as rural vs. urban schools.
In addition to location and type of school, school-going pupils may be exposed to sexual abuse in different parts or spaces in the larger school environment. A study by Ruto (2009) reported that teacher's houses were the riskiest spaces with 53% of sexual abuse incidents followed by other parts of the school compound (36%) and lastly bushes around the school (2%). Similarly, in the USA, classrooms, lavatories, teacher's quarters, and dormitories were showed to be risky spaces for learners (Claudia, 2010) and so are sports time and sports arena (Mat, 2016). A clear understanding of spaces that predispose children to sexual abuse is needed for any meaningfully interventions to be put in place.
Reports on child sexual abuse globally show that girls carry a larger burden of this pernicious challenge to public health. However, gaps in knowledge exist on the magnitude of the condition that is borne by boys especially among primary school-going children. It can be argued that a holistic understanding of the distribution patterning of its burden by type and location of schools is needed to enable formulation of targeted interventions.

Study Population
The study targeted both male and female school-going children between the ages of 10 -16 years who were randomly selected from the speci ed strata of Kisumu County primary schools.

Methods of data collection
Questionnaires were administered to pupils whose parents had consented and who individually assented to participate in the study. Self-administered questionnaires were used as they provided anonymity given the sensitive nature of the questions (Sussana G and Jaap D, 2016). In order to further guarantee anonymity and con dentiality, the pupils were asked to drop the answered questionnaire into a sealed box. In addition, focus group discussions (FGDs) and Key Informant Interviews (KII) were conducted to collect data on pupils' and key persons who are in touch with pupils (guiding and counselling teachers) perceptions, attitude, and knowledge about sexual abuse, as well as on spaces both within and outside the school that predisposed children to sexual abuse. A pilot study, using 10% of the sample size in schools not selected to participate in the study.

Data analysis
All data collected was coded based on the variable and response types and analysed as per the objectives.
Institutional names were coded and were only accessible to the principal researcher. Chi-square tests for proportions were used to determine association between type and sexual abuse among school-going children. Descriptive statistics were used to analyse frequency data on spaces that predispose school going children to child sexual abuse. Qualitative data were analysed thematically.

Ethical considerations
Study was reviewed and approved by Maseno University Ethics Review Committee. Administrative approval from the Ministry of Education of Kisumu County was then sought to facilitate undertaking of the study in primary schools as they oversee all schools. Head teachers were also asked for permission to conduct the study in their respective schools. The schools and pupils participated voluntarily without fear of victimization; this was achieved by giving parental consent forms to all pupils who wished to participate in the study among the selected classes for uniformity then from those whose parents consented, the pupils were asked for assent independently and voluntarily, among those who accented they were randomly selected. Consent and assent was sought from the parents and pupils who were assured of both privacy and con dentiality. Con dentiality was achieved through coding of the questionnaires and pupils did not write any of their personal identi cation details anywhere. They were then educated on the nature and purpose of the study and how they will be needed to answer the questionnaire. More so, con dentiality was achieved through archiving the data in a password protected computer. Interviews were conducted within secluded areas to ensure both audio-and visual privacy. Pupils were not interviewed in presence of their peers, parents or teachers. However, the respective school Guidance and Counselling teachers were requested to be on standby to help in case of need of counseling support to minimize psychological harm.

Results
There were 53% (n= 212) girls and 47% (n= 185) boys. The mean age (SD) for both boys and girls was 13 years (SD= 1.584) years with girls dominating. There were signi cantly more pupils in public day schools than boarding ones (p= 0.017; df= 1; X 2 = 5.742) ( Table 1). Among the pupils who had faced any type of sexual abuse either by touch, verbally or visually, more girls than boys had experienced any type of sexual abuse, majority being by 'sight' (65%) compared to boys, who had experienced sexual abuse more frequently through 'touch', 41%. Chi-square analysis summarized in Table 2 show there was signi cant gender experience of verbal sexual abuse (p < 0.014).
Twenty-six percent of the pupils who responded they had experienced sexual abuse (by touch, verbally or visually) were in public schools while 20% were in private schools but the difference was not statistically signi cant (x 2 = 2.044; p>0.05). Table 3. Qualitative information from the focus group discussions however, revealed that pupils generally thought that public day schools exposed school-going children to sexual abuse compared to boarding or private schools. Table 4 shows pupil's views.
Twenty-six percent of the pupils who indicated that they had experienced sexual abuse (touch, verbal or visual) were in rural schools compared to 21% in urban schools (x 2 = 0.823, p>0.05). From the focussed group discussions, some pupils perceived that, attending rural schools entailed higher risk for sexual abuse than in the urban ones, but there was no consensus as shown in table 5.
Pupils who indicated that they had experienced child sexual abuse (irrespective of gender, type or location of school) were 25%. Table 6 shows some of the places pupils pointed out as the sites where abuse happened to them. It is notable that school was the most frequent mentioned site where pupils experienced abuse from. Bushes inside school, latrines and bushes outside school were the frequently cited spaces where pupils perceived sexual abuse takes place within the school across both genders at 52%, 51% and 50% respectively as shown in table 7.

Discussion
More girls (53%) than boys (47%) reported that they had experienced sexual abuse. Indeed, the disproportionate burden of sexual abuse experienced by girls is re ected in the perspectives of the schoolgoing children. For instance, some of them felt that girls are at more risk or face the highest burden of sexual abuse because girls are perceived as weak and lack the ability to defend themselves. More so, that they (girls) were easily emotionally blackmailed thereby making them more susceptible to being victims of child sexual abuse. 5.1 Association between type of school and forms of sexual abuse among school-going children In this study, whereas there was no statistically signi cant in uence of type of school on the type of sexual abuse (by touch, verbally or visually) girls and boys had experienced, the risks remain real. Qualitative information however, revealed lack of consensus on pupils' perception of predisposition to sexual abuse by type of school. A section of the pupils perceived attending public day schools increased the possibility that one would experience sexual abuse given the predisposing factors are more pronounced for these type of schools such as: walking long distances to and from school; odd times for reporting to and exiting school that is occasioned by preparatory or remedial classes; congestion in the public day schools; lack of adequate security; lax school rules; and poor conduct of some school teachers. On the other hand, other pupils perceived being in boarding schools more likely predisposed one to sexual abuse as they were in a con ned environment than their counterparts in day schools, and while they can experience abuse, they may lack somebody to tell like a parent due to many inhibitions. Whether this nding remains merely a subjective nding was not proven statistically. Consequently, even though the United Nations Children's Fund (UNICEF, 2006) has recommended that manual transit between school and home should be treated as school environment, evidence from this study suggests that holistic interventions are needed-those that cover safeguards both at the school, during pupil's commute, and at home.
While cases of sexual abuse in private schools were lower than those in public schools, pupils reported that they lacked someone to report to or open up to with such delicate issues. Previous studies have shown that cases of sexual abuse that take place in elite schools are infrequently reported as parents of these pupils pay high tuition fees and exposing such a vice in their schools, taints the name of the school (Warkov, 2017). However, private school pupils were in an environment that had good security and strict rules, and that any report of abuse is non-tolerable that may contribute to lower occurrences of child sexual abuse.
Public school pupils might be more predisposed to child sexual abuse for reasons such as: teachers didn't care a lot about what happens to pupils, theirs is only to teach, that public schools are big geographically hence more spaces and opportunities for hiding, which together expose pupils to sexual abuse as cited by pupils. In addition, respondents also noted that public schools are congested and many do not have adequate security.

Association between location of school and sexual abuse among school-going children
Pupils who came from rural areas were perceived to face higher risk for sexual abuse than those from the urban areas indicating there might be indeed disproportionate risk burden of this vice in the rural areas.
Some reasons cited by pupils and teachers why rural pupils were more predisposed to child sexual abuse than those in the urban included: schools are in bushy areas and people in the rural areas were ignorant on child sexual abuse. Vastness of rural schools created more space under which child sexual abuse predators would commit the vice. While other studies (Shumba et al., 2015) may have documented that rural school pupils are more vulnerable to child sexual abuse, the current study found no such association. According to the ecological model, determinants of sexual violence is the result of complex interplay of multi-level factors (inherent within and external to the individual) and it is plausible that risk factors interact or cluster differentially across contexts, hence the different observations in literature.

Spaces that predispose school-going children to sexual abuse
The current study in which primary schools were surveyed in Kisumu Central and Kisumu West sub-Counties, a number of risky spaces were cited by pupils as predisposing to sexual violence, but at home (38%) and at school (36%) were predominant while the 'way home' was cited as a potential risky space. Whereas there were diverse reasons and hot spots for child sexual abuse happening at school or between home and school, it is important to point out that a considerable proportion of the respondents had experienced sexual abuse in their homes. However, the study did not determine if there were pupils who had ever experienced repeat sexual victimization/abuse. While previous studies have pointed to this possibility and that of co-occurring types of sexual abuse (Meinck et al, 2016), there is hardly any literature on this aspect in Kenya, thus an aspect for future research.

Conclusions
1. Female pupils are more likely than males to experience any form of sexual abuse. Also, visual sexual abuse, which is a subtle form, was more predominant.
2. The study concludes that there is still a fuzzy relationship between type of school and child sexual abuse where pupils attending public day schools were perceived to be more likely to experience abuse due to congestion in schools and security issues while accessing school and even within the school itself whereas there was no statistical evidence for this in the current study.
3. The study did not establish an association between where a school is located and child sexual abuse.
Furthermore, while the architectural plan in most rural schools allows large spaces of playground and room for growth, this may present increased risks to child sexual abuse for pupils. Sites that are isolated and those with minimal supervision increase exposure of pupils to sexual abuse.

Declarations
My study will not brew any con ict of interest. I was the principle investigator and I independently coordinated my research team (research assistants) during data collection. I did all the data entry and analysis. It was a self-funded research. My university supervisors have been recognized as co-authors as they helped in the development of the nal document.
The study followed all laid ethical rules in the country.
World Bank (2005). The situation faced by young people in kenya.     the other side and they use one path going to class, when they use the path, coz sometimes they study at night and night is dark so they will just take the advantage and go to the corridors and maybe attack the girls'    Table 5