Rwanda - Agricultural Household Survey 2017

In order to provide the basis for conducting sample surveys based on complete coverage of the household level in all 30 districts of Rwanda, and as a better way of collecting agricultural household data and finding better precise survey estimates, agricultural household survey(AHS) used a Multiple-Frame Sampling (MFS) methodology by which, area frame was constructed and survey sample was drawn from it.
In the sampling strategy of the SAS 2017, it was proposed that 960 PSUs be selected in the first instance by systematic sampling method with probability proportional to size. At the second stage the sampled PSUs were divided into SSUs among which only one SSUs was sampled at random for each PSU and used for the survey. 960 segments were drawn from three agricultural strata including intensive agriculture land on hillsides (stratum 1.1), intensive agriculture land in marshlands (stratum 2.0), rangelands (stratum 3.0). Using the open segment obliges to include urban strata in the sampling frame. Apart from stated strata, village stratum which combines two substrata; urban area (stratum 4.1) rural settlements (stratum 4.2) was added to the sample frame. This stratum was divided into segments and a sample of 600 segments was drawn by systematic sampling method. It should be mentioned that the Stratum 4.1 and 4.2 were specially added to the purpose of having a complete sampling frame for estimation of livestock since a major portion of livestock are associated with households located in villages and rural urban area adjacent to grazing land that had previously been missing for all of the previous SAS area sampling frames.
This survey was conducted among private agricultural households, in 4 strata used in Seasonal agricultural survey. Households were listed from the 1,560 sampled segments in the country. Among the listed households 16,057 households were found the ones applying agricultural activities. A single visit was done in every identified agricultural household and the information collected covered two agricultural seasons A and B of 2017 agricultural year.



Taking into consideration of full completeness of the segment, sampling weights were calculated for each stratum in each district considering the total number of segments in the stratum and the sample size in the specific stratum. The weight was calculated as the inverse of the overall probability of selection, taking into account the probabilities from all sampling stages. For strata 1.1, 2.0 and 30, the sampling of segments was done through two stages and probabilities were calculated for each sampling stage. For strata 41and 42 the sampling of segments was done through one stage and probabilities were calculated in order to determine the weight.