All households were selected randomly by either selection of the case (if sampled from registration frame) or selection of the building. In instances, in which a sampled building contains more than one household, one randomly selected household within the building is interviewed. All individuals within the household are further selected randomly and therefore represent respective underlying populations. 4,050 households, of which 2,300 were refugee households, 650 former refugees, and 1,200 households from the host community, were targeted based on expected credibility of results. A strict replacement and sample protocol were used to address the non-response or ineligibility of sampled households. The replacement sample was usually drawn together with the main sample, and selection into a replacement sample was randomized and strictly controlled. Replacements were made in cases of ineligibility, i.e., refugee households in host community sample as well as in cases of non-response. The final realised sample of the FDS in Zambia was 4,310 households. Among the sample, 2,373 are refugee households, 722 former refugee households, and 1,215 are host community households. The higher sample realization compared to the target is a result of the addition of additional out of settlement households interviewed in Lusaka but registered in Mayukwayukwa.
