Key Social Norms Take-aways The COMPASS intervention in Ethiopian refugee camps influenced social norms around gender, education, marriage, and social support, even though it did not significantly reduce violence within the 12-month evaluation period. Girls in the intervention reported more gender-equitable beliefs regarding education, marriage age, and childbearing, and had stronger social networks. These shifts reflect early movement along the social norms and empowerment pathway necessary to prevent violence, particularly in complex humanitarian settings where normative change often precedes behavior change.
Objectives The study aimed to evaluate whether the COMPASS life skills and safe spaces intervention could reduce experiences of interpersonal violence among refugee adolescent girls in Ethiopia, and to assess its impact on social norms, support networks, and empowerment indicators.
Methods A cluster randomized controlled trial was conducted with 919 Sudanese and South Sudanese adolescent girls aged 13–19 in three refugee camps in Ethiopia. Participants were divided into 62 clusters and randomly assigned to intervention or control groups. Quantitative data were collected at baseline and 12-month follow-up, using self-administered surveys through audio computer-assisted self-interviewing. Analyses included mixed effects logistic and linear regressions, with adjustments for age, relationship history, and parental cohabitation.
Intervention Details The COMPASS intervention included 30 weekly group life skills sessions in safe spaces for adolescent girls, facilitated by trained female refugee mentors. Sessions covered communication, friendship-building, gender-based violence, and reproductive health. In parallel, caregivers participated in eight monthly discussion groups addressing adolescent development and violence prevention. The intervention also provided access to unstructured safe spaces for girls.
Results The intervention did not significantly reduce girls’ experiences of sexual, physical, or emotional violence within the one-year timeframe. However, girls in the intervention arm showed significantly improved attitudes about delaying marriage and childbearing, increasing school completion, and reported stronger social connections, including having friends and trusted non-family female adults. Among girls married at baseline, the intervention was associated with a reduction in child marriage at follow-up.
Conclusions Although the COMPASS intervention did not reduce violence exposure in the short term, it successfully shifted social norms and increased social support for girls. These intermediate outcomes are critical steps along the pathway to violence prevention. Longer-term follow-up and broader inclusion of community and male stakeholders may be necessary to achieve reductions in violence in humanitarian settings.