RWANDA - Comprehensive Food Security and Vulnerability Analysis and Nutrition Survey 2009

Rwanda is administratively divided into four provinces (Northern Province, Southern Province, Eastern Province and Western Province) plus Kigali City and a total of 30 districts. Districts are further divided in sectors and cells. The 2009 CFSVA and Nutrition Survey was designed to provide statistically representative information at the sub-provincial level. To facilitate comparison with existing studies, it was decided to define strata using administrative limits rather than food economy zones (as in 2006). Because of the large number of districts, it was decided to define strata that would be either single districts or a group of districts. Districts that were identified as similar with regards to their socio-economic and agroenvironmental characteristics were grouped together. A total of 16 strata were defined including 8 districts and 8 groups of districts. Kigali City was not included in the sample.

The Eastern Province includes the strata of Nyagatare-Gatsibo-Kayonza, Kirehe-Ngoma-Rwamagana, and Bugesera.
The Northern Province includes the strata of Musanze-Burera, Gakenke, and Rulindo-Gicumbi.
The Western Province includes the strata of Rubavu, Nyabihu, Ngororero, Rutsiro-Karongi, and Nyamasheke-Rusizi.
The Southern Province includes the strata of Kamonyi-Muhanga-Ruhango, Nyanza, Huye, Gisagara, and Nyamagabe-Nyaruguru.

Within each stratum, NISR implemented a two-stage sampling procedure to select households using an approach that is standardized for statistical studies in Rwanda. Zones de Dénombrement (ZD, enumeration areas) were selected first, followed by households using 2007 population estimates based on the 2002 census. The ZDs are a sampling unit that is smaller than a sector.
A total of 450 ZD were selected. In each stratum, the probability of the ZDs to be selected was equal to the number of ZDs in the stratum divided by the number of ZDs. In each stratum, ZDs were randomly selected. Within each sampled ZD, a total of 12 households were interviewed, resulting in a total expected sample size of 5,400 households.

All of the households were interviewed. Enumerators were provided with clear instructions on which households to interview, and how to find them. Supervisors were provided with a list of over-sampled households in the event that a household had to be replaced.

Because this study also focuses on the relation between nutrition and food security, it was decided during the study design that only households with children aged below 5 years old would be included in the sample. This imposed some limitations in the ability to draw conclusions about all the households in Rwanda, as explained in the limitations section below. (See in CFSV&NS 2009 Report)


Taking into consideration the sampling methodology summarized above, adjustment weights were computed to provide results representative at country level. The household probability of selection is equal to the product of a household’s probability of being selected in a ZD by the probability of the ZD of being sampled.
The inverse of this probability is the design weight. The design weight is divided by the product of the total number of households in the population divided by the number of sampled households. The result is the normalized weights which were used in all analyses.