Stormwater Management and Climate Change Adaptation Impact Evaluation 2014

Detailed maps of the study area have been drawn up in consultation with the local partners. After discussions with ADM and the local NGOs, it was decided that the most appropriate unit of intervention would be the quartier, or neighborhood. The intervention zone is divided into two phases, phase 1 where PROGEP infrastructure investments are currently ongoing, and phase 2, where the start date of these investments is dependent on additional project financing. In total, 395 quartiers are recorded in the entire zone, 143 in phase 1 and 252 in phase 2. Within this area, 160 quartiers are to be included in the study, with 80 quartiers in the treatment group and 80 in the control group. In each quartier, a focal CBO is in the process of being identified and a sample of 2,400 households will be randomly selected across all quartiers included in the study (15 per quartier). Quartiers are grouped according to a set of criteria drawn up by the social facilitators, who have intimate knowledge of the study areas, on the basis of geography and social ties between quartiers. There are a total of 47 groupings or sites, 27 in phase 1 and 20 in phase 2. Within each grouping, quartiers are randomly assigned to treatment and control groups. This will take place by way of public lottery to ensure transparency and buy-in from the local communities involved.

As the results for Phase 1 and Phase 2 areas may be quite different, they will be analyzed separately. Therefore, to give an upper bound for the minimum detectable standard effect size (MDES), the relevant number of sites is 20. Table 1 below gives MDEs for household level indicators for different ICCs and variation explained by co-variates (R2), assuming 20 sites and 4 quartiers per site, 2 treatment and 2 control. The number of households to be surveyed per quartier is 15. All calculations are done with power at 80% and significance level at 5%.