Violence Against Children and Associated Factors Among High School Students in Jimma Town

Experiencing violence in childhood impacts lifelong health and well-being. Yet being subject to violence at a young age causes enduring psychological, social and neurological vandalism which stops people reaching their full potential in life. Violence against children includes all forms of violence against people under 18 years old ; therefore the purpose of this study is to determine the magnitude of violence against children and associated factors among sampled high school students in Jimma town. Method: Institution based crossectional study was employed among sampled 423 private and public high school students of Jimma town from february15-30, 2018. Self-administered structured questionnaire was used to collect the data. A multi stage with proportional to sample size was employed. Result: A total of 419 respondents participated in the study, making the response rate 99%. A large proportion of the study subjects, 57.5 %, were females, and 56.3% were Oromo’s. The study indicates that the prevalence of violence among high school children’s’ in Jimma town is 62.5%. Children who had faced severe physical violence accounted for 14.1%. Pertaining to parental relationships and interactions, the majority of the school children (66.3%) did not have free discussion on reproductive health issues with their parents or guardians. Conclusion: The prevalence of any form of violence in life time (Physical, sexual and psychological) found to be high. Of all children who experienced violence near to half of them encountered mild physical violence. Less number of school children had experience of Kchat chewing and/or alcohol use.


Methods And Participants Study area and Design
The study has been conducted in Jimma Town public and private secondary (high) schools from february15-30, 2018. Jimma is located 356 km south-west of Addis Ababa. The town has estimated population of 192,000, and there are four public secondary schools (Abba Buna, Jiren, Saxoo, and Jimma University Community), and six private secondary schools (Eldan, Catholic, Beteseb, Tesfa Tewahido, Ebifam, and Mowada). Jimma is the largest town in south-western Ethiopia, situated in Oromia Region in Jimma Zone. The study has been conducted in selected public and private secondary schools. The study has employed school based cross sectional study design. All public and private secondary schools found in Jimma Town comprise the source population. The study population consists of three public and two private schools from Jimma Town.
Secondary school students whose age is under 18 years were included since they are considered as children in Ethiopia, considering proportional number of girls and boys from each school and grade.
Sample size determination and sampling procedure Using single population proportion formula, assuming 50% prevalence of violence against children, 95% con dence level, and 5% marginal error and by adding 10% of non-response rate, the sample size was 423. The survey has been conducted on randomly selected two private and three public secondary schools. A multi stage with proportional to sample size was employed considering type of school, students school grade and class section. From each selected section, students were randomly (lottery method) taken using students registration book by considering sex proportion.
Data collection tools Self-administered structured questionnaire was used to collect the data which had ve parts such as sociodemographic characteristics of the child and parents, prevalence with forms of violence, child's history of substance uses and the attributes of family behaviors, perpetrators of violence against children and related variables, and existing prevention mechanisms.The questionnaire was initially prepared in English and then translated into two local languages such as Amharic and Afaan Oromo. The questionnaire passed through a three-phase review process. In the first phase, teachers and school staff reviewed the items; in phase 2, the research team tested the items on 5% a sample in primary schools in Seka Town to check the understanding of the students and finally, the modi ed version of the questionnaire were duplicated and administered to 423 secondary school students in each selected sections within the class room.

Method of Data Analysis
Data were checked for completeness, consistency and entered into Epi Data version 3.1. SPSS version 21.0 was used for statistical analysis after cleaning. A logistic regression model was used to identify explanatory variables. Candidate variables at p-value<0.25, in bivariate analysis, were entered into multiple logistic regressions. Binary logistic regression analysis used to see the values of COR which was declared as signi cant at p-value < 0.05. Backward model selection method was used. The degree of association between dependent and independent variables were assessed using an adjusted OR with 95% CI. The Hosmer and Lemeshow test were used to check model tness at P-value of 0.05.

Result And Discussion
Socio-demographic characteristics A total of 419 respondents participated in the study, making the response rate 99%. A large proportion of the study subjects, 57.5 %, were females, and 56.3% were Oromo's. Nearly half, 50.4% of the study participants were followers of Orthodox Christianity. In terms of residence, almost of the respondents, 93.6%, were from Jimma Town. Similarly, most of the study subjects (92.6%) fell in the age range of 15-19 years while the rest (7.4%) fell in the younger age category of 10-14 years. In terms of educational status, 53.9 % of the respondents were attending grade 9 while 46.1 % were from grade 10. Most of the respondents (93.6%) were residents of Jimma Town, whereas only 6.4 came from outside the town. Findings related to educational performance indicate that 31.5 % had very good result followed by who had good result, accounting for 21.7 %. Respondents with satisfactory and poor educational performance accounted for 17.7 and 3.6 % respectively. Regarding marital status or sexual relationships, the study has shown that a large majority (72.1%) were single, followed by those students who had boyfriend or girlfriend, accounting for ( 22.4 %). Married students and co-ha biting ones accounted for 4.1 and 1.4 % respectively. The other important socio-demographic characteristics investigated were parental relationships and status. Accordingly, large majorities (77.8%) were from married and living together family, while respondents from broken families due to divorce and deceased partner accounted for 11.2% and 8.15% respectively. Respondents with orphan status accounted for 2.9 %. In terms of parental educational status, the ndings show that a large majority (82.6%) had literate father, while only 6.2 % had illiterate father. Among the literate fathers, 38.25 had attained above grade 12. Those who attained grades 11 to 12, 9 to 10 and 1 to 8 accounted for 22.8, 22.5, and 16. 5 % respectively. Similarly, a large majority (82.6%) had literate mothers whereas only 8.1 % had illiterate mothers. Among the literate mothers, a large proportion (46.5%) attained grades 9 to 10, followed by those who attained grades 1 to 8, accounting for 21.4 %. Those mothers who attained above grade 12 and 11 to 12 grades accounted for 20.5 and 11.6 % respectively.
Regarding provision of money for basic needs, a large majority (73.7%) reported that they got it, while 26.3 % reported lack of such support. (See Table 1) The prevalence of violence against children The study indicates that 262 or 62.5% of the school children living in Jimma Town had experienced some form of violence in their life; 12.9%, 7.9%, 46.5% encountered mild sexual violence, severe sexual violence and mild physical violence respectively. Children who had faced severe physical violence accounted for 14.1%.
Respondents' who had encountered emotional or psychological violence accounted for 37%. (See Table 2) Children history of substance use and the attributes of family behaviors Pertaining to parental relationships and interactions, the majority of the school children (66.3%) did not have free discussion on reproductive health issues with their parents or guardians. As to the level of family control, a little bit more than half of the respondents (51.6%), 42% and 6.4% experienced tight, moderate and loose family control. The majority (72.6%) did not witness parental con ict at home. With regard to substance use, the study has shown that most of the school children (88.8%) did not chew Chat. Similarly, most of the respondents (94%) did not smoke cigarettes at all. A large majority 82.3%) had never consumed alcohol while 17.7% of the respondents had already used alcohol. The use of illegal drugs such as hashish and marijuana was even more restricted among school children as 95.2% did not use it at all. (Table 3) Perpetrators of Violence against Children and Related Variables Alarmingly enough, the three highest perpetrators of physical violence against children were found to be boyfriend/girlfriend, teachers, and family members, accounting for 28.3, 23.3 and 22.2 respectively. With regard to emotional or psychological violence, the leading perpetrators were teachers, boyfriend/girlfriend and family members, accounting for 38%, 21.4, and 16.7 respectively. Similarly, the highest perpetrators of sexual violence were boyfriend/girlfriend, teachers, and family members, accounting for 35.3%, 25.9% and 14.4% respectively. A large proportion of the violence against children (41%) occurred during the daytime, while 28 % was perpetrated at night. Regarding spots of violence, school, streets and home account for 34.5%, 32.5% and 24%, respectively. The majority of school children (66.5%) put the blame on the perpetrators of violence, while 15% and 14.5% took the blame themselves and attributed it to the community, respectively. Gender difference was not considered important in exposing children to violence by 58% of the respondents. Regarding the interpersonal relationships between the school children and their parents/guardians, 58.9% had positive interactions most of the time; 64.7% always accepted and followed their advice and guidance, while 16.9% sometimes rejected them. The majority (65.2%) always had positive and smooth interactions with their school teachers; only 9.8% had negative and rocky relationships. (See Table 4)

Existing prevention mechanisms of violence
In terms of awareness, the majority (69.2%) claimed to have known some mechanisms that would protect them from violence. The three most important mechanisms included being disciplined and respectful to others; knowing where, with whom and when to move; and knowing the behavior of other person before establishing close relationships,accounting for 39.3%, 25.9%, and 23.8%, respectively. As rst res ponders,48.4% mentioned family members, while any person in the nearby and friends were mentioned by 18.1% and 15.5%, respectively.
A large proportion (57.3%) depended on strict rules and regulations to be protected from violence at school, followed by parents and teacher's committee, mentioned by 20% of the respondents. In residential areas, 49.9%, 25.1% and 21.7% sought guarantee from patrolling police, community leaders and Kebele militia force, respectively. A large majority (81.6%) indicated that they had obtained advice from their parents or guardians as to how they could protect themselves from violence. (See table 5) Factors Associated with Violence First, bi variate logistic regressions were done to identify candidate variables. From this step, 14 variables were subjected to be candidate using P-value of <0.25. Those were parental relationship (P-0.005), gender (P-0.076), child's achievement (P-0.116), mother's educational status (P-0.171), father's educational status (P-0.001), witnessing parental con ict (P-0.001), family control (P-0.050), Ever Kchat chewing (P-0.001), ever drunk alcohol (P-0.000), interaction with parents/guardian (P-0.001), reaction to parents advice (P-0.003), interaction with teachers (P-0.131), money gift from family (P-0.001), and educational status of the child (grade) (P-0.176). Out of these candidates, only ve variables were become independent and signi cantly associated factors that affect occurrence of violence among school children. Those were fathers' educational status (P-0.021), money gift from family (P-0.005), witnessing parental con ict (P-0.007), ever drunk any alcohol (P-0.034), and interaction with parents (P-0.006).
In the multi-variable logistic regression analysis; father's educational status, money gift from the family, witnessing family con ict, ever drunk alcohol and interaction with parents were independent predictors of violence against children. The odds of being experiencing violence among children borne from illiterate fathers was three times more likely than children born from literate fathers [AOR=2.827, 95% CI: (1.173, 6.814)].
Children who did not have access to money as a gift from family were two times more likely to be exposed for  Table 6)

Discussion
Violence against children remains a pervasive, but largely ignored issue in many parts of the world, particularly in Africa. This is certainly the case in Ethiopia, where children regularly face humiliating physical punishment, sexual violence and psychological abuse at home, in school and in the community-at-large. This study assessed the prevalence and associated factors of violence among 419 high school students in Jimma Town, Ethiopia. The prevalence of any form of violence (Physical, sexual and psychological) found to be 62.5%. This nding is consistent with studies conducted in Ambo and Jimma, and Northern Nigeria (9)(10)(11) .
This high level of prevalence might be due to the sensitivity of our measuring item which was "did you face any form of violence in your life?" This includes any type/forms of violence encountered in life and very wide to accommodate from mild to severe violence in each forms.
Among the students who experienced violence, 12.9%, 7.9%, 46.5% encountered mild sexual violence (rape), severe sexual violence (rape) and mild physical violence respectively. This nding is lower than studies conducted in other African countries; the life time prevalence of sexual violence (rape) in South West Nigeria and Urban Zimbabwe was 42.1% and 33% respectively (12,13) .
This discrepancy might be due to social and cultural variation between the study subjects in reporting sexual violence or rape. In Ethiopian context, sexual related offence particularly rape is a strong cultural norm including abduction. Moreover; the chance of engaging in marriage is considered as minimal for the female rape survivor afterwards. Surprisingly, there were also male children faced sexual violence that indicates there is a practice of homosexuality. However, they might keep secret for fear of the stigma and other negative responses from the community. From those points of view, under reporting might be in our context in both sex.
Pertaining to parental relationships and interactions, the majority of the school children (66.3%) did not have free discussion on reproductive health issues with their parents or guardians. As to the level of family control, a little bit more than half of the respondents (51.6%), 42% and 6.4% experienced tight, moderate and loose family control. Tighter control over the children without free discussion will be one of the pushing factors that could expose them for different forms of violence. Moreover, less disclosure of RH issues may be associated with lack of social support, and hence higher likelihood of violence will happen. In addition, witnessing family violence may increase the likelihood of later victimization to any forms of violence especially for sexual violence. This could be explained by learning and acceptance of the victim's role of the early abusive environment (14)(15) .
With regard to substance use, the study has shown less number of school children (11.2%) had experience of kchat chewing. Less than one fth (17.7%) of the respondents had already used alcohol. This nding is lower than the magnitude reported from Jimma Zone Limu Genet district. This might be due to children under 18 years only included to our study. They may consider it as sensitive issue and may not provide their actual experience; student's willingness in giving genuine information might underestimate the magnitude of the problem and social desirability bias may also affect it (6,16) .
The common perpetrators of physical violence against children were found to be boyfriend/girlfriend, teachers, and family members, accounting for 28.3, 23.3 and 22.2 respectively. With regard to emotional or psychological violence, the leading perpetrators were teachers, boyfriend/girlfriend and family members, accounting for 38%, 21.4, and 16.7 respectively. Similarly, the highest perpetrators of sexual violence were boyfriend/girlfriend, teachers, and family members, accounting for 35.3%, 25.9% and 14.4% respectively. This nding is consistent with studies conducted among school students in Addis Ababa, Jimma, and Tanzania that reported greater number of respondents were perpetrated by someone they closely knew and the most common perpetrators were found to be friends, teachers, peers and close relatives (17)(18)(19)(20) .
Near to half of the respondents encountered violence (41%) that occurred during the daytime, from which 34.55 happened at school compound. The majority of school children (66.5%) put the blame on the perpetrators of violence, and some other attributed it to the community. Gender difference was not considered as an important issue in exposing children to violence by 58% of the respondents and gender difference is not signi cant predictor at the nal regression model too. This nding is in line with what had been reported from else where (21) . Regarding the interpersonal relationships between the children and their parents/guardians, 58.9% had positive interactions most of the time; 64.7% always accepted and followed their parents' advice and guidance, the majority (65.2%) always had positive and smooth interactions with their school teachers. This nding may have its own impact or implication on reducing the life time prevalence of violence in our study, without this good interaction, it could have been more than that. Different proportions for each category of interaction type were not compared due to of scarcity of published article.
In terms of awareness to existing prevention mechanisms, majority (69.2%) claimed to have known some ways that would protect them from violence. The three most important mechanisms included; being disciplined and respectful to others; knowing where, with whom and when to move; and knowing the behavior of other person before establishing close relationships, accounting for 39.3%, 25.9%, and 23.8%, respectively. This is some what comparable with others nding. This may be due to the norm or cultural similarities on discipline and respectful behavior among study participants in different areas of Jimma especially during young age (6) .
However, beside those mechanism that would protect them from violence, a large proportion (57.3%) depended on strict rules and regulations to be protected from violence at school. In residential areas, 49.9% sought guarantee from patrolling police. A large majority (81.6%) indicated that they had obtained advice from their parents or guardians as to how they could protect themselves from violence. Therefore, we can say that even if there were many protective factors, still children were under violence in different forms with high prevalence.
Predictors identi ed from nal regression model such as educational status or child grade level, money gift from family, witnessing parental con ict, drinking any type of alcohol, and type of interaction with parents were consistently reported in other studies. However, lack of free discussion with parents and gender difference were strong predictors of violence elsewhere but not in our study. This might be due to paternalistic approach of child caring in Ethiopia considered as a norm for both sex (9,11,19) .

Conclusion
The prevalence of any form of violence in life time (Physical, sexual and psychological) found to be high. Of all children who experienced violence near to half of them encountered mild physical violence. Pertaining to parental relationships and interactions, the majority of the school children did not have free discussion on reproductive health issues with their parents or guardians. With regard to substance use, the study has shown less number of school children had experience of Kchat chewing and/or alcohol use. The common perpetrators of physical violence against children were found to be boyfriend/girlfriend. A large proportion of the violence against children occurred during the daytime since high proportion of physical violence reported. Regarding place of violence, the school areas take the highest. Regarding the interpersonal relationships between the children and their parents/guardians and school teachers, majority had positive interactions in most of the time; always accepted and followed their parents' advice and guidance, always had positive and smooth interactions with their school teachers. In terms of awareness, more than three-fourth claimed to have known some mechanisms that would protect them from violence. Finally, we recommend that considering predictor identi ed from this study would be paramount to design intervention among school children to minimize violence. Ethical clearance and an approval letter were obtained from Jimma University institute of health-institutional reviewing board. Study participants were informed about the objective and details of the study including publication. Informed written consent obtained from each subject, and assent was obtained from their parent or care giver or legal guardian to participate in the study. Con dentiality was maintained by using anonymous codes. Ethical letter and consent form can be provided up on your request.

Consent to publish
All parties involved (Jimma University-the funding organization study subjects and authors) agreed to publish on international peer reviewed journal. During data collection, all participants informed and agreed on the major objective of the study which is for academic purpose including publication. Permission obtained for publication available with the corresponding author.

Availability of data and materials
The data sets used and/or analyzed during the current study are available from the corresponding author on reasonable request. The SPSS (software) which is completed with raw data set can be also shared. All data generated or analyzed during this study are included in this manuscript.

Competing/Con ict of interest
All authors declare that they have no any nancial and non-nancial competing interests. None of the authors of this paper has a nancial or personal relationship with other people or organizations that could inappropriately in uence or bias the content of the paper. It is to speci cally state that "no competing interests are at stake and there is no con ict of interest" with other people or organizations that could inappropriately in uence or bias the content of the paper.

Funding
The research was nancially supported by Jimma University Institute of health.

Authors' contribution
All authors had made equal and substantial contributions to conception and design, acquisition of data, analysis and interpretation of data, prepared the manuscript critically for important intellectual content, and worked together starting from proposal development to write up. All authors read and approved the nal manuscript.