Estimates of Sampling Error
The sample of respondents selected in the Thailand Multiple Indicator Cluster Survey (MICS) is only one of the samples that could have been selected from the same population, using the same design and size. Each of these samples would yield results that differ somewhat from the results of the actual sample selected. Sampling errors are a measure of the variability between all possible samples. The extent of variability is not known exactly, but can be estimated statistically from the survey results.
The following sampling error measures are presented in this appendix to the Survey Final Report (attached in the External Resources) for each of the selected indicators:
- Standard error (se): Sampling errors are usually measured in terms of standard errors for particular indicators (means, proportions etc). Standard error is the square root of the variance. The Taylor linearization method is used for the estimation of standard errors.
- Coefficient of variation (se/r) is the ratio of the standard error to the value of the indicator
- Design effect (deff) is the ratio of the actual variance of an indicator, under the sampling method used in the survey, to the variance calculated under the assumption of simple random sampling. The square root of the design effect (deft) is used to show the efficiency of the sample design. A deft value of 1.0 indicates that the sample design is as efficient as a simple random sample, while a deft value above 1.0 indicates the increase in the standard error due to the use of a more complex sample design.
- Confidence limits are calculated to show the interval within which the true value for the population can be reasonably assumed to fall. For any given statistic calculated from the survey, the value of that statistics will fall within a range of plus or minus two times the standard error (p + 2.se or p - 2.se) of the statistic in 95 percent of all possible samples of identical size and design.
For the calculation of sampling errors from MICS data, SPSS (Statistical Package for Social Sciences) Version 14 Complex Samples module has been used. In addition to the sampling error measures described above, the tables also include weighted and un-weighted counts of denominators for each indicator.
Sampling errors are calculated for indicators of primary interest, for the national total, for the regions, and for urban and rural areas. Three of the selected indicators are based on households, 8 are based on household members, 13 are based on women, and 15 are based on children under 5. All indicators presented in the Final Report are in the form of proportions.
Other forms of Data Appraisal
A series of data quality tables and graphs are available to review the quality of the data and include the following:
Age distribution of the household population
Age distribution of eligible women and interviewed women
Age distribution of eligible children and children for whom the mother or caretaker was interviewed
Age distribution of children under age 5 by 3 month groups
Age and period ratios at boundaries of eligibility
Percent of observations with missing information on selected variables
Presence of mother inthe household and person interviewed for the under 5 questionnaire
School attendance by single year age
Sex ratio at birth among children ever born, surviving and dead by age of respondent
Distribution of women by time since last birth
Scatterplot of weight by height, weight by age and height by age
Graph of male and female population by single years of age
The Thailand Multiple Indicator Cluster Survey (MICS) covered a large number of samples from all 76 provinces in the country. It was expected that data deviation could possibly occur from the work of the field staff, or the interviewees. Therefore, the National Statistical Office (NSO) operated a post enumeration survey (PES) in Bangkok and 22 provinces selected from all four regions to aid data users in their consideration of data quality. The PES consisted of 150 block/village samples, in both municipal and non-municipal areas. Collective household samples – 20 households per block/village for a total of 3,000 household samples – were selected from the listing of household samples of the MICS survey. Staff were sent in to repeat the survey in these areas. Matching of questionnaires from the actual survey and the repeated survey was carried out and data were analysed for deviation.