Mongolia - MICS 2010

The primary objective of the sample design for the Mongolia Multiple Indicator Cluster Survey 2010 was to produce statistically reliable estimates of most indicators, at the national level, for urban and rural areas, and for the regions (Western, Khangai, Central, Eastern, and Ulaanbaatar) of the country. Urban and rural areas in each of the five regions were defined as the sampling strata. A two-stage, stratified cluster sampling approach was used for the selection of the survey sample.

Sample Size and Sample Allocation.

The target sample size for the Mongolia MICS 2010 was calculated as 10500 households at the national level, 2100 households at the regional level. For the calculation of the sample size, the key indicator used was the pre-school attendance among children aged 3-4.

From the results of Mongolia MICS 2005, the pre-school attendance among children aged 3-4 was 37 percent at the national level while it is 32 percent in Western region, 32 percent in Khangai region, 33 percent in Central region, 40 percent in Eastern region, and 48 percent in Ulaanbaatar. The value of deff was calculated as 1.7 at the national level, 1.2 in Western region, 2.0 in Khangai region, 2.5 in Central region, 2.0 in Eastern region, and 1.3 in Ulaanbaatar. Also, from the 2009 annual statistics on population, the percentage of children aged 3-4 in the total population was 4.4 and average household size was 4. The resulting number of households from this exercise was, at the beginning, 1572 households for Western region, 2472 region for Khangai region, 3156 households for Central region, 1839 households for Eastern region and 873 households for Ulaanbaatar, - thus yielding 10183 households in total. The number households for each region vary greatly between regions, from 873 to 3156. Therefore, equal allocation of the total sample size to the regions is used in order to keep sampling errors at similar level for regions. As a result, the number of households needs to be selected from each region rounded up to 2100 households and then the total sample size is determined as 10500 households.

The average number of households selected per cluster for the survey was determined as 25 households, based on a number of considerations, including the design effect, the budget available, and the time that would be needed per team to complete one cluster. Dividing the number of households to be selected from each region by the number of sample households per cluster, it was calculated that 84 sample clusters would need to be selected in each region - thus yielding 420 clusters in total. As mentioned above, equal allocation of the total sample size to the five regions was used. Therefore, 84 clusters were allocated to each region, with the final sample size calculated at 10500 households (84 clusters *5 regions * 25 sample households per cluster). In each region, the clusters (primary sampling units) were distributed to urban and rural domains, proportional to the size of urban and rural populations in that region.

Sampling Frame and Selection of Clusters.

The 2009 annual statistics on population and households frame was used and the khesegs of khoroos of Ulaanbaatar and baghs of soums of provinces are defined as clusters. As first stage of the sampling, the clusters were selected from each of the sampling strata by using systematic pps (probability proportional to size) sampling procedures, based on the estimated sizes of the enumeration areas from the 2009 annual statistics on population and households.

Household Listing and Selection.

The Statistics Divisions in provinces and districts were responsible for asking the governors of selected baghs and khesegs (PSUs) to update their household listings and for sending the updated listings to the National Statistics Office (NSO). The governors of the selected baghs and khesegs were instructed to include all households locating in the territory of the bagh or kheseg regardless of their registration. As second stage of the sampling, the households were then sequentially numbered from 1 to n (the total number of households in each cluster) at the NSO, where the selection of 25 households in each cluster was carried out using random systematic selection procedures.

The major features of the sample design are provided in the paragraph Sampling of External Resources. Sample design features include sampling stages and stratification, target sample size and its allocation, sampling frame and selection of clusters, household listing and selection, and the calculation of sample weights.


No major deviations from the original sample design were made. All primary sampling units were accessed and successfully interviewed with good response rates.


In total 10,500 households selected for the sample, of these 10,092 households were successfully interviewed for a household response rate of 98 per cent. Of the listed household members in the interviewed households, 91 per cent of women aged 15-49, 96 per cent of mothers and caretakers of children under age of 5, 78 per cent of men aged 15-54, and 97 per cent of mothers and caretakers of children aged 2-14 successfully interviewed.

The survey covers data from 35,981 household members from 10,092 households. Of the households interviewed, 6,086 or 60 per cent are urban and 4,006 or 40 per cent are rural. About half of these households have 3-4 persons, households with 1-2 members account for 26 per cent, and those with more than 5 members - 24 per cent. The mean household size is 3.2 persons. 22 per cent of households are female headed.

Of surveyed 35,981 respondents, the sex ratio was 96 males to 100 females. Because of increased fertility rates from 2006, children aged 0-4 constitute 11 per cent of the total population. 61 per cent of total population is men aged 15-59 and women aged 15-54. Figure 2 presents age and sex distribution of household population.

By marital status, 64 per cent of the women are currently married or in a union, 26 per cent are never been married or in union, 4 per cent are divorced, 3 per cent widowed and 3 per cent are separated. 19 per cent of total women had given a birth to a child in the last two years.

By education level, 3 per cent of females have no education, 5 per cent are primary level educated, 10 per cent with vocational education, 19 per cent have lower secondary education, 29 per cent have upper secondary education and 34 per cent have college, university education.

From total of 3,956 children under 5 covered by the survey, male proportion is 50 per cent and female proportion is 50per cent. By education level of the mothers/ caretakers, 5 per cent have no education, 7 per cent have vocational education, 8 per cent are primary educated, 19 per cent are lower secondary educated, 27 per cent with upper secondary education, and 36 per cent have college, university education. The distribution of children under age of 5 by the wealth index of the households shows that, 24 per cent live in lowest quintile, 20 per cent in second quintile, 19 per cent in middle quintile and 19 per cent in fourth quintile and the rest 18 per cent in highest quintile.

65 per cent of all surveyed men are married or in a union, 30 per cent are never been married or in union, and the remaining 5 per cent are either divorced, separated or widowed. Males have lower level of education compared to females, 6 per cent have no education, 9 per cent have primary education, 10 per cent have vocational education, 23per cent have college, university education, 26 per cent lower secondary education, and 26 per cent have upper secondary education.

The sex ratio of total 9,131 children aged 2-14 covered by the survey is 105 boys per 100 girls. Four per cent of the mothers/ caretakers have no education, 8 per cent have primary education, 11 per cent have vocational education, 21 per cent have lower secondary education, 27 per cent with upper secondary education and 29 per cent have college, university education.


Calculation of Sample Weights

The Mongolia Multiple Indicator Cluster Survey 2010 sample is not self-weighting. Essentially, by allocating equal numbers of households to each of the regions, different sampling fractions were used in each region since the size of the regions varied. For this reason, sample weights were calculated and these were used in the subsequent analyses of the survey data.

The major component of the weight is the reciprocal of the sampling fraction employed in selecting the number of sample households in that particular sampling stratum (h) and PSU(i).

A second component in the calculation of sample weights takes into account the level of non-response for the household and individual interviews.

Finally, the design weights were calculated by multiplying the above factors for each for target group and cluster. These weights were then standardized (or normalized), one purpose of which is to make the weighted sum of the interviewed sample units equal the total sample size at the national level. As a result, the range of the normalized weights calculated for each target group is shown below and these sample weights were appended to all data sets and analyses were performed by weighting each household, woman aged 15-49, under-5, men aged 15-54, and child aged 2-14 with these sample weights.

- Households - 0.145-3.876
- Women aged 15-49 - 0.138-3.794
- Children under-5 - 0.142-4.042
- Men aged 15-54 - 0.139-3.712
- Children aged 2-14 - 0.149-4.238
Detailed information and formulas for calculation of the sampling procedure are provided in the External Resources of this documentation.