Community Action to End 'Early Sex' in Kenya: Endline Report on Community-led Child Protection

EXECUTIVE SUMMARY

Community-level child protection has long been a priority in both humanitarian and development settings. Until recently, however, there has been little research using robust methodologies to analyze the effectiveness of community-level child protection interventions. A 2009 global, inter-agency review found a weak evidence base for the effectiveness of Child Welfare Committees, one of the most widely used child protection interventions at that time. The evidence indicated that Child Welfare Committees achieved low levels of community ownership, had modest levels of effectiveness, and were unsustainable. Local people tended to see them as ‘NGO projects’ and depended on the NGOs for continuing them.

Accordingly, an inter-agency meeting decided to develop and test systematically the effectiveness of more community owned processes of child protection that link with formal, government aspects of child protection, and to use the learning from the research to strengthen practice. Overseeing the process was an Interagency Learning Initiative on Community-Based Child Protection Mechanisms and Child Protection Systems, coordinated by Save the Children (London). The technical arm of the initiative—the Child Resilience Alliance—agreed to lead the technical aspects of the research and decided to use a participatory action research approach. Kenya and Sierra Leone were selected as the sites for the action research. This report is the final evaluation or endline report on the action research in Kenya.

Action Research Design and Stages

In Kenya, the action research was conducted in Kilifi County in Coast Province in areas populated by Giriama people. The study used a two-arm cluster randomized trial design in Marafa and Bamba—two approximately matched areas. The Marafa location (comprised of two adjoining villages constituting one community) was randomly assigned to the intervention arm while the Bamba location (comprised of two adjoining villages constituting one community) was assigned to the comparison arm. This design enables one to make causal attributions regarding the effectiveness of the intervention.

The initial phase of the action research consisted of rapid ethnography, which examined questions such as who is a child and what are the main harms to children. In both Marafa and Bamba, leading harms to children were lack of food, being out of school, early pregnancy, overwork, drug abuse, poor parenting, and bad behavior by children. Sexual exploitation and abuse of girls was widespread, as girls who were hungry took food from men, who demanded sex in return. Also, girls took rides from ‘boda boda’ (motorbike taxi drivers), who then demanded sex as a form of payment. Teenagers and young people took part in disco matangas, funeral celebrations that raised money to help pay families’ funeral expenses but were sites of mass drinking and sexual abuse of girls by men. Many girls became pregnant, dropped out of school, and more than a few got married at a young age.

Following the ethnographic phase, a baseline survey was used to collect data in both the intervention and comparison villages regarding children’s risks, protective factors, and well- being in July, 2016. Next, the intervention villages in Marafa engaged in a slow, inclusive process of dialogue to decide which harm or harms to children to address subsequently through a community-led action. Concerned about the pervasiveness of early sex, including girls as young as 8 years, and related issues of sexual exploitation, early pregnancy, early marriage, and school dropout, the communities themselves decided to address early sex. After the villages had selected early sex as the issue to be addressed and had outlined their implementation plan, a second, short baseline survey was used to collect information focused specifically on early sex and related problems such as sexual exploitation, teenage pregnancy, and school dropout.

In Marafa, the intervention communities addressed early sex through their own self-designed and collectively implemented action (2017-2019). The community-led action included: community dialogues about the importance of avoiding engagement in early sex; girls and boys playing football as a means of avoiding idling and engaging in sexual activity; learning life skills, often in discussions associated with football practice or games; encouragement of girls to stay in school; community theater and dialogues to raise awareness of the problems of early sex and how to prevent it; discussions between parents and girls about how to prevent early sex and other problems; and the community successfully petitioning the Chief to ban disco matanga.

Endline Methodology

The endline study, which was conducted October-November, 2019, used a mixture of quantitative and qualitative methods with a target population of children aged 10-17 years in Marafa and Bamba. The survey used the same questions that had been used in the baseline surveys, and was conducted in Giriama by nine trained Kenyan researchers backstopped by experienced Kenyan and international researchers. Survey data were collected using smart phones. Differences between the intervention and comparison communities were analyzed using Analysis of Variance (ANOVA) and nonparametric tests, with statistical significance evaluated at the p<.05 level. Particular attention was given to interaction effects, which, if significant, would indicate whether the intervention had a significant effect while taking into account baseline-endline changes in the comparison condition.

Qualitative methods such as in-depth interviews enabled learning from the narratives and lived experiences of the children (10 – 17 years) and caregivers in both Marafa and Bamba. Group discussions with different sub-groups made it possible to contrast the perspectives of groups such as girls and boys or children and adults. The questions explored changes that had occurred in the last two years, and also inquired about the community-led intervention itself. The questions were not asked in a structured or semi-structured manner since the intent was to follow the partcipants’ line of thought and explore topics they thought were important. The qualitative data were collected in Giriama and translated into a verbatim, English, written transcript that included no names or personal identifiers. The data were analyzed using a grounded methodology that helped to identify natural categories consistent patterns. The research ethics had been reviewed and approved by both Pwani University in Kenya and the Institutional Review Board of Randolph-Macon College.

A limitation of the study is that it does not involve a nationally representative sample. Hence, it is important to avoid overgeneralizing the findings.

Key Findings

Girls and boys in Marafa became key actors, decision-makers, and influencers in the community-led action. They were the ones who chose football plus discussions as part of the intervention. Children were key in influencing peers, talking about the importance of staying in school, sending messages about avoiding early sex by means of community theater, and setting positive role models. Another important process result was that the people in Marafa took ownership for the community-led action to address early sex. Relying upon themselves and including a diversity of actors, they saw the community-led action as their own, not as an ‘NGO project’. Natural helpers, who wanted to help children, were key resources and led the conceptualization and implementation of the community-led action, without pay.

Early sex. Turning next to the outcomes for children, early sex decreased significantly in Marafa. Among girls and boys in the age range of 8 – 11 years, the decrease in Marafa was greater than the modest decrease that had occurred in Bamba, where a youth group had been active around issues such as teenage pregnancy. Similarly, early sex was reduced among girls 12-15 years, with the reduction being greater than that which had occurred in Bamba. Concommitantly, the average age at which girls began engaging in sex increased in Marafa from a mean of 12.7 years (baseline) to 14.6 years (endline). In Bamba, the mean age increased from 12.4 years (baseline) to 13.7 years (endline). Both girls and boys in Marafa frequently attributed the reductions in early sex to their involvement in football activities and the accompanying guidance and life skills. Girls commented on the value of developing life skills such as the ability to say ‘No” to men. In contrast, girls in Bamba reported that the risks of early sex, early pregnancy, and early marriage remained strong.

Teenage pregnancy. The narratives of girls, boys, and adults indicated that teenage pregnancy had decreased significantly in Marafa due to support from parents, life skills such as saying ‘No’, staying in school, positive role modeling, and men’s awareness of how lack of basic necessities such as sanitary towels made girls vulnerable to men. Children and adults agreed that the principal factor in reducing teenage pregnancy in Marafa was that local people had banned together and petitioned the Chief to ban disco matanga. The Chief of Marafa confirmed that ending disco matanga played a role in reducing early pregnancy.

In contrast, people in Bamba reported consistently that early pregnancy was a significant problem and that 8 girls out of 30 from the primary school in Bamba had become pregnant in the school year in which the endline data were collected. Girls frequently became pregnant due to having unmet needs for items such as sanitary towels. Boda boda and even grown men exploited girls for whom they had provided such items. In addition, HIV in children was reportedly a widespread problem in Bamba, though it was not discussed openly.

Improved Parental Care of Children. In Marafa, parents worked together with the advice of a teacher to talk with their children about puberty, sex, and pregnancy. They also learned how to set rules in the home regarding, for example, treating each other with respect, and the importance of monitoring their children and knowing their location and activities. The parents reported that they enjoyed talking with their children and had begun to treat them better. The children, too, said consistently that they enjoyed being able to talk with their parents in ways that they had not done previously. Parents also showed increased commitment to meeting girls’ basic needs as a way of reducing transactional sex. In Bamba, no such activation occurred for parents around caring for their children. Indeed, women complained that the men were disinterested and alcohol abusers.

School Participation. In Marafa, the reduction in idling and early sex, together with changing peer and parental influences on children, led to reduced school dropout from primary school and increased participation and learning in school. The parents attributed increased school participation to the improved parenting and to the community formation of an out of school committee that responded to cases of out of school children. Bamba showed no increase in school participation, and no parents and civic groups working to limit school dropout. Sometimes being out of school led children to engage in activities such as disco matanga that led to high rates of pregnancy.

Spread of the intervention. A positive yet unexpected development was that the children and the parents Marafa helped to spread the community-led intervention to neighboring villages. Football plus discussions, including at tournaments, served as the primary means through which the spread occurred. In discussions, people from the neighboring villages asked whether they could have the community-led intervention in their own villages. Parents, too, reached out to neighboring villages, sharing their learning and accomplishments. In Bamba, there were limited positive supports for children, no dynamic ambassadors, and no spread to neighboring villages.

Linkage with Government services. The action research worked closely with the Department of Children’s Services via its Field Office. K. Ondoro was frequently invited by the Office to give inputs on different issues or to give trainings to Government child protection officers in different Counties.

Implications for Action

Collectively, these findings have significant implications for community-level child protection work and efforts to strengthen child protection systems, both in Kenya and internationally. Practitioners, donors, and policy leaders should:

• Make wider use of community-led approaches to child protection, which are effective in reducing early sex, including the sexual exploitation of girls.
• Appreciate that community-led child protection is a natural means of implementing an ecological, relational approach.
• Support community resilience and self-reliance by using community-led approaches, which are locally owned and sustainable.
• Enable children’s leadership, going beyond adult driven approaches, unlocking children’s agency, and positioning children as agents of change on behalf of their communities.
• Prioritize prevention, using community-led approaches to complement to top-down, responsive work that is an important part of child protection.