Sampling
Sampling Procedure
The 2016 HSES used the sampling frame which was developed by the NSO based on 2015 population figures obtaine from administrative records. The design of the survey recognizes two explicit strata that urban (Ulaanbaatar, aimag centers), and urban (sum centers and the countryside). The Primary Sampling Unit (PSU) was kheseg in Ulaanbaatar and bags in aimag center and rural areas, and 1836 PSUs covered in survey whole year. In order to comparison of results by aimags, sample was allocated proportionally and PSUs selected two stages of simple random sampling. In Ulaanbaatar, 360 PSUs, 24 PSUs from each aimag (36 PSUs in Darkhan-Uul, 60 PSUs in Orkhon and 12 PSU s in Govisumber aimags) were selected, while for the rural, 48 PSUs (60 PSUs in Uvurkhangai, 60 in Khuvsgul, 24 in Darkhan-Uul, 0 in Orkhon and 24 PSUs in Govisumber aimags). Then 10 households selected from each PSU in urban areas and 8 households in each PSU in rural areas. The sample of 16488 households was allocated as follows: 3,600 in Ulaanbaatar, 5400 in aimag centers and 7488 in rural areas and sum centers. However, the actual sample size used for this analysis is slightly smaller: 3573 households in Ulaanbaatar; 5394 in aimag centers; and 7484 in rural areas. In order to obtain representative
statistics for each stratum and for the country as a whole, it was necessary to use sampling weights.
