Heterogeneous impacts of interventions aiming to delay girls' marriage and pregnancy across girls' backgrounds and social contexts
Objectives
The paper aimed to examine how the impacts of interventions designed to delay girls’ marriage and pregnancy vary by girls’ backgrounds and social context in Bangladesh and Zambia. Using data from two randomized controlled trials, it assessed whether program effects differ according to girls’ literacy, household wealth, village-level girls’ paid work participation in Bangladesh, and community-level prevalence of premarital sex and sexual violence in Zambia.
Findings
In Bangladesh, academic skill training significantly reduced child marriage in villages where girls’ paid work participation, especially in acceptable “white-collar” jobs, was relatively high, while gender-awareness training was more effective in villages with low paid work participation, particularly for illiterate girls. In low-participation villages, all three arms, and especially the livelihood arm, were effective among girls from middle-wealth households, suggesting that acceptable work opportunities and agency interact with household status in shaping impacts. In Zambia, empowerment clubs (“safe spaces”) combined with health vouchers significantly reduced early pregnancy in communities with high premarital sex prevalence, with stronger effects among illiterate girls, and safe spaces reduced early pregnancy in communities with high sexual violence regardless of literacy, whereas impacts were negligible in low-risk communities. PIIS1054139X21004602
Recommendations
The authors recommend tailoring child marriage and pregnancy prevention programs to specific social contexts and to the most vulnerable girls. In settings where marriage decisions are collective and girls’ work options are constrained, interventions should strengthen girls’ negotiation skills and expand socially acceptable educational and livelihood opportunities. In contexts where premarital sex and sexual violence are common and pregnancy may precede marriage, priority should go to empowering illiterate and at-risk girls through safe spaces linked to health services. They argue that program designers should use a socioecological lens, systematically consider community norms and opportunities, and build evidence on which combinations of components work best for different groups and contexts.
Share your research
You can share details of your ongoing and upcoming research to be included in the CRANKs online research tracker. By doing this, you are contributing to a coordinated, harmonised global research agenda.