Survey ID Number
MMR_2009_MICS_v01_M
Title
Multiple Indicator Cluster Survey 2009-2010
Sampling Procedure
The primary objective of the sample design for the Myanmar Multiple Indicator Cluster Survey was to produce statistically reliable estimates of most indicators, at the national level, for urban and rural areas, and for the 17 states and divisions of the country: Kachin, Kayah, Kayin, Chin, Mon, Rakhine, Shan (North), Shan (East), Shan (South), Ayeyarwaddy, Bago (East), Bago (West), Magwe, Mandalay, Sagaing, Tanintharyi, and Yangon.
Although aiming to produce data for the national level, 10 townships in Ayeyarwaddy Division and four townships in Yangon Division affected by Cyclone Nargis in 2008 were excluded from the sampling frame. Moreover, five townships in Shan (North) State were removed from the sampling frame due to security concerns. Urban and rural areas in each state/division were defined as the sampling domains. A multi-stage, stratified probability proportional to size (PPS) cluster sampling design was adopted for the selection of the survey sample.
The target sample size for the Myanmar MICS was calculated as 29,250 households.
The resulting number of households from this exercise was 29,250 households. The average cluster size in the Myanmar MICS was determined as 30 households, based on a number of considerations, including the budget available, and the time that would be needed per team to complete one cluster. Dividing the total number of households by the number of households per cluster, it was calculated that the selection of a total number of 975 clusters would be needed for the Myanmar MICS.
A compromise allocation procedure was adopted for allocation of 975 clusters to the 17 states and divisions. In each state/division, the clusters (primary sampling units) were distributed to urban and rural domains, proportional to the size of urban and rural populations in that state/division.
The 2006 updated list of villages from the Population Department, Ministry of Population and Immigration was used as the sampling frame for the selection of clusters. In urban areas wards were defined as primary sampling units (PSUs), and in rural areas village tracts were defined as PSUs. Before sampling, wards and village tracts of 19 townships in Ayeyarwaddy, Yangon and Shan (North) were removed from the sampling frame. PSUs were selected from each of the sampling domains by using systematic PPS (probability proportional to size) sampling procedures. Within each village tract, one village was selected by simple random sampling.
During the fieldwork period 40 of the selected enumeration areas were not visited because they were considered inaccessible due to security concerns. These were replaced with other clusters of similar size. Substitution of selected clusters is, however, not a recommended MICS procedure. According to MICS standard protocol when sampled enumeration areas are not accessible, they should be abandoned without selecting replacement clusters. This is because the replacement clusters will not have been selected with the same probability of selection as the other clusters. The situation in these geographical areas might also systematically differ from the areas which were not possible to visit, which could introduce bias in the data.
Since the sampling frame (the 2006 list of villages from the Population Department) was not up to date, the number of households in all selected clusters was updated in consultation with local village/ward authorities prior to the selection of households. For this purpose, survey teams visited each enumeration area and listed the occupied households one day ahead of field data collection. Survey teams were required to either obtain a map from the local authorities, or draw a map themselves, detailing all dwellings in the enumeration area. The map was then divided into segments of around 15 households. Two segments were selected at random by the field team, to make up a cluster of 30 households.
The sampling procedures are fully described in "Myanmar Multiple Indicator Cluster Survey 2009-2010 - Final Report" pp.124-126.