Health Results Based Financing Impact Evaluation 2015, Household Baseline Survey

The household sampling follows a two-stage strategy. Enumeration areas (EAs) were initially defined by the 2006 census but in case the enumeration area covered two different catchment areas, the EA was split into smaller units defined by the catchment areas. First, EAs within the catchment area of each health center were ranked for listing by random drawing. In each of the villages in the EA ranked first, a dedicated listing team did a household listing, collecting information on the presence of women who were pregnant in the preceding two years and of adolescent females. The listing team did a complete listing whether eligible or not. In cases in which the EA had less than 80 households or if less than 18 eligible households for any of the two samples were listed, the next ranked EA was added. Listing was always done at the EA level, i.e. all households in the EA were listed unless a village was not found by the listing team.
Second, up to 50 households were then randomly selected by the research team for survey. The selection of households was according to the following two criteria: (a) 25 households with at least one birth or pregnancy, regardless of the outcome of the pregnancy, in the two years preceding the survey; (b) 25 households with at least one adolescent girl. In cases with less than 25 eligible households listed, all eligible households were selected for the survey. The selection of the ‘recent pregnancy’ and of the ‘adolescent’ sample was done separately. However, in some cases, the same household was randomly selected for both samples.

The final household sample includes 2120 households. From these the research team was able to link 1252 female questionnaires and 972 adolescent questionnaires. Deleted from the raw sample were 41 female questionnaires and 9 adolescent questionnaires that could not be matched to a household roster.