Node 1. Maps of the six study areas/clusters are available and they are of the resolution and the detail that would be required for the study. These maps are taken from most recent simple spatial survey (S3M) conducted by the FMoH alongside UNICEF. A map of Kassala state in which the six study clusters are identified along with programme sites and villages was created and can be found in Figure 1. The same map data can be used to create the detailed maps for each study area/cluster.
Figure 1: Detailed map of Kassala state with study areas/clusters and programme sites identified
Node 2. This approach is based on centric systematic area sampling (CSAS) methods1. Detailed instructions on how to segment each study area/cluster using CSAS methods can be found here.
Node 3. This is done by listing out all the villages in each of the segments and then randomly choosing a village from this list for example by lottery. It should be noted that this step is to be done for baseline and for each of the steps of the study for a total of 5 times. It would therefore be advisable that when the list of villages per segment is done during baseline, this list should be kept in a spreadsheet and/or in a database so that the selection of villages for the next steps can be much easier and quicker. Random sampling can also be facilitated as spreadsheet tools can be used to perform the random selection.
Node 4 and Node 5. For baseline and at each step of the study, a full list of selected villages for sampling should be put together. These lists should be provided to respective survey teams. The list will also help in preparing for the survey round (e.g., planning movement within the locality to cover all sampling villages, sending advance notice to village leaders regarding survey, etc.)
Node 6 and Node 7. In the occasion that there are no detailed high resolution maps available, list of villages per study areas/clusters can be used for stage 1 sampling. The list of villages should be organised by sub-locality to be able to ensure that the ensuing sample is as much as possible spatially even. Using the list, systematic sampling is performed. This is done by first determining a sampling interval and second determining a random starting point from which to begin the systematic selection. The sampling interval is calculated by dividing the total number of villages by the number of villages that are needed to sample. Detailed instructions on how to perform a stage 1 systematic sampling with a random start can be found here.