Water treatment facilities are critical in rural Rwanda to ensure safe drinking water. However, their efficiency varies significantly. A quasi-experimental design was employed with random assignment of villages to treatment or control groups. Data on water quality and usage patterns were collected over one year. Random assignment led to a statistically significant Y = β₀ + β₁X + u where Y represents the treated village's improved water quality, X is the intervention (treatment), with robust standard errors of ±5% suggesting moderate effect sizes. The replication confirms the initial findings and provides methodological insights for future evaluations in similar contexts. Further trials should consider additional factors such as socio-economic status and environmental conditions to enhance generalizability.
Ngirumwami et al. (Thu,) studied this question.