Adolescent girls in Kenya who are marginalized face considerable risks and vulnerabilities that affect their educational status, health, and general well-being.

  • They are at high risk for early marriage, unintended pregnancy, early and unprotected sex, sexual assault, and HIV and other sexually transmitted infections.
  • They have limited income-earning opportunities and high rates of illiteracy, and often experience violence and social isolation.
  • In addition to all these challenges, they are frequently living in the context of acute poverty at the household and community levels.

Council research has shown that it is important to reach these girls when they are young—between the ages of 10 and 14—before irreversible events like early and/or unintended pregnancy can anchor them in poverty. Evidence shows that there is potential that if girls are reached in early adolescence, prior to negative outcomes setting in, they can be set on a more positive life trajectory.

The Adolescent Girls Initiative-Kenya (AGI-K) tests layered packages of interventions that reflect the complex challenges and underlying cultural barriers that adolescent girls in Kenya face. AGI-K is a randomized controlled trial implemented in two sites in urban Kibera, an informal settlement in Nairobi; and in rural Wajir, a remote, underdeveloped area in Kenya along the Northeast border with Somalia.

AGI-K reached more than 6,000 girls ages 11–15 over two years.

AGI-K tests the following four layered packages:

  • Violence Prevention, which includes community dialogues on the prevalence and persistence of violence against girls, coupled with funded community projects to address these challenges.
  • Violence Prevention + Education, which also includes a cash transfer conditional on school enrollment at the start of each term and regular attendance throughout the term.
  • Violence Prevention + Education + Health, in which girls meet in safe space groups once a week with a female mentor who delivers a health, life skills, and nutrition curriculum.
  • Violence Prevention + Education + Health + Wealth Creation, in which girls also participate in additional financial education sessions integrated into the “safe space” curricula and open savings accounts or home banks.

At the end of the two-year program, AGI-K midline results show positive impacts for girls across a broad range of health, social, educational, and financial indicators in both Kibera and Wajir. Individual results varied across both sites.

  • In Kibera, AGI-K had significant effects on violence reduction, primary school completion and schooling self-efficacy, sexual and reproductive health knowledge, social safety nets, financial literacy and savings behavior, and household economic status.
  • In Wajir, AGI-K had statistically significant effects on primary school enrollment, positive gender norms and self-efficacy, financial literacy, and savings behavior.

Endline results confirm that AGI-K’s approach is protective, especially for girls who are most marginalized. Findings reinforce that conditional cash transfers are a key driver of improvement in education outcomes, and of delaying sexual debut and pregnancy as well as marriage and pregnancy.

Furthermore, results show that layering multiple interventions to holistically address girls’ complex needs can increase the overall positive impact on girls’ lives. This analysis points to the win-win opportunity to maximize girls’ welfare by introducing education, health, and wealth interventions together—both lowering costs and increasing benefits for girls.

Information from randomized controlled trials can be reliably used to shape the field of adolescent girls’ policy and programs. This project identifies best practices, refines the critical elements of girl-centered programs, and helps to eliminate ineffective approaches.

The AGI-K intervention ended in 2017. The study’s midline assessment results were released in 2018 and endline results of the program’s longer-term impacts were published in 2020. The Council will build on this evidence—testing the most effective models for scaling up in partnership with the Wajir County Government.