Multiple Indicator Cluster Survey 2010 - Bhutan

Abstract
---------------------------

The National Statistics Bureau conducted the Bhutan Multiple Indicator Survey between March and August, 2010. The survey’s main objective is to provide up-to-date information on the situation of children and women in Bhutan. The survey is also aimed at furnishing data required for monitoring progress towards the MDGs, the goals of A World Fit for Children and other international goals. It is hoped that the findings will serve as a basis for equity-based programming, as well as contribute towards the improvement of data and monitoring systems in Bhutan. It will also help to strengthen technical expertise in the design, implementation, and data analysis of similar surveys in future.

Geographic coverage
---------------------------

Three regions and 20 Dzongkhags (districts)

The Western region includes Chhukha, Gasa, Haa, Paro, Punakha, Samtse and Thimphu Dzongkhag; the Eastern region includes Lhuentse, Monggar, Pemagatshel, Samdrup-Jongkhar, Trashigang and Trashiyangtse Dzongkhag and the Central region includes Bumthang, Dagana, Sarpang, Trongsa, Tsirang, Wangdue Phodrang and Zhemgang Dzongkhag.

Analysis unit
---------------------------

- individuals
- households

Universe
---------------------------

The survey covered all de jure household members (usual residents), all women aged between 15-49 years, all children under 5 living in the household.

Kind of data
---------------------------

Sample survey data [ssd]

Sampling procedure
---------------------------

The primary objective of the sample design for the Bhutan Multiple Indicator Survey was to produce statistically reliable estimates of most indicators at the national level, for urban and rural areas, and for the 20 Dzongkhags of the country. Urban and rural areas in each of the 20 Dzongkhags were defined as the sampling strata.

A multi-stage, stratified cluster sampling approach was used for the selection of the survey sample.

The target sample size for the BMIS was calculated as 15,400 households. For the calculation of the sample size, the key indicator used was the stunting among children aged 0-4 years.

The resulting number of households from this exercise was 800 households, which is the sample size needed in each Dzongkhag except Gasa (200 households) after taking account of the finite population correction factor - thus yielding about 15,400 in total. Gasa was a special case, in that it has a very small population, and widely dispersed. It was felt that 200 households was the maximum sample size that could realistically be achieved in that Dzongkhag.The average number of households selected per cluster for the BMIS was determined as 20 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 total number of households by the number of sample households per cluster, it was calculated that 40 sample clusters would need to be selected in each Dzongkhag except Gasa (10).

Equal allocation of the total sample size to the 20 Dzongkhags was used except Gasa. Therefore, 40 clusters were allocated to each Dzongkhag except Gasa (10), with the final sample size calculated at 15,400 households: 40 clusters * 19 Dzongkhags * 20 sample households per cluster and 10 clusters*1 Dzongkhag *20 sample households per cluster in Gasa. In each Dzongkhag, the clusters (primary sampling units) were distributed to urban and rural domains, proportional to the size of urban and rural populations in that Dzongkhag. The table below shows the allocation of clusters to the sampling strata.

The sampling procedures are more fully described in "Bhutan Multiple Indicator Cluster Survey 2010 - Report" pp.207-212.

Mode of data collection
---------------------------

Face-to-face [f2f]

Research instrument
---------------------------

The questionnaires for the Generic MICS were structured questionnaires based on the MICS4 model questionnaire with some modifications and additions. Household questionnaires were administered to a knowledgeable adult living in the household. The household questionnaire includes Household Listing Form, Education, Water and Sanitation, Household Characteristics, Child Labor, Disability, and Hand Washing.

In addition to a household questionnaire, questionnaires were administered in each household for women age 15-49 and children under age five. The questionnaire for children under 5 years of age was administered to mothers or caregivers of all children under 5 years of age living in the households.

The women's questionnaire includes Woman's Background, Child Mortality, Desire for Last Birth, Maternal and Newborn Health, Contraception, Unmet Needs, Marriage/Union, Attitudes towards and Experience of Domestic Violence, Sexual Behaviour, HIV/AIDS and Maternal Mortality.

The children's questionnaire includes Child's Age, Birth Registration, Early Childhood Development, Breastfeeding, Care of Illness, and Anthropometry.

Cleaning operations
---------------------------

Data was entered using the CSPro software in 25 micro-computers and the entry was carried out by 25 operators and three supervisors. In order to ensure quality control, all questionnaires were double entered and internal consistency checks were performed. Procedures and standard programs developed under the global MICS4 programme and adapted to the Bhutan questionnaire were used throughout. Data entry began a month after the start of data collection and was completed in September 2010. Data was analysed using the Predictive Analytics Software (PASW), the version 18 of SPSS software, and the model syntax and tabulation plans developed by UNICEF were used for this purpose.

Response rate
---------------------------

Of the 15,400 households selected for the sample, 14,917 were occupied. Of which, 14,676 households were successfully interviewed for a household response rate of 98.4 percent. Within those interviewed households, 16,823 of the eligible women (aged 15-49) were identified. Of them 14,018 were successfully interviewed, yielding a response rate of 83.3 percent. The household interviews identified 6,457 children under-five. The questionnaires were completed for 6,297 of them with a response rate of 97.5 percent.

Sampling error estimates
---------------------------

Sampling errors are a measure of the variability between the estimates from all possible samples. The extent of variability is not known exactly, but can be estimated statistically from the survey data.

The following sampling error measures are presented in this appendix 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 of the estimate. 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, and is a measure of the relative sampling error.
• 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 in relation to the precision. 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, with a specified level of confidence. For any given statistic calculated from the survey, the value of that statistic will fall within a range of plus or minus two times the standard error (r + 2.se or r – 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 Version 18 Complex Samples module has been used. The results are shown in the tables that follow. In addition to the sampling error measures described above, the tables also include weighted and unweighted counts of denominators for each indicator.

Sampling errors are calculated for indicators of primary interest, for the national level, for the regions, and for urban and rural areas. Eight of the selected indicators are based household members, eighteen are based on women, and twelve are based on children under-five. All indicators presented here are in the form of proportions.

Data appraisal
---------------------------

A series of data quality tables are available to review the quality of the data and include the following:

- Age distribution of the household population
- Age distribution of eligible and interviewed women
- Age distribution of children under 5 in household and children under 5 questionnaires
- Women’s completion rates by socio-economic characteristics of households
- Completion rates for under-five questionnaires by socio-economic characteristics of households
- Completeness of reporting
- Completeness of information for anthropometric indicators
- Heaping in anthropometric measurements
- Observation of places for hand washing
- Observation of women's health cards
- Observation of children under 5 birth certificates
- Presence of mother in the household and the person interviewed for the under-5 questionnaire
- School attendance by single age
- Sex ratio at birth among children ever born and living

The results of each of these data quality tables are shown in appendix D in document "Bhutan Multiple Indicator Cluster Survey 2010 - Final Report" pp.240-251.

摘要
---------------------------
国家统计局于2010年3月至8月期间进行了不丹多指标调查。调查的主要目标是提供关于不丹儿童和妇女现状的最新信息。调查还旨在提供监测MDGs（千年发展目标）、《适合儿童的世界》目标以及其他国际目标进展所需的数据。希望研究结果能够作为基于公平的规划的基础，并有助于提升不丹数据与监控系统。同时，也有助于加强未来类似调查在设计、实施和数据分析方面的技术专长。
地理覆盖范围
---------------------------
调查涵盖了三个地区和20个宗（区）。西部区域包括楚克哈、加萨、哈、帕罗、彭措、桑特西和蒂姆 Pu Dzongkhag；东部区域包括卢恩特、蒙加尔、佩马加特舍尔、桑杜普-琼卡、特拉希冈和特拉希扬特塞 Dzongkhag；中部区域包括布姆唐、达嘎纳、萨尔庞、通萨、西朗、旺都波德朗和泽芒 Dzongkhag。
分析单元
---------------------------
- 个人
- 家庭
总体
---------------------------
调查涵盖了所有法定家庭成员（常住居民），所有15-49岁的女性，以及所有居住在家庭中的5岁以下儿童。
数据类型
---------------------------
样本调查数据 [ssd]
抽样程序
---------------------------
不丹多指标调查的样本设计的主要目标是产生对大多数指标在国家层面、城市和农村地区，以及国家20个宗的统计上可靠的估计。在每个20个宗的城市和农村地区被定义为抽样层。
采用多阶段、分层集群抽样方法进行抽样样本的选择。
BMIS的目标样本量计算为15,400户。在计算样本量时，使用的关键指标是0-4岁儿童的生长迟缓。
此活动产生的家庭数量为800户，这是在考虑有限人口修正系数后每个宗所需的样本量，因此总计约15,400户。加萨是一个特殊情况，因为它人口非常少，分布广泛。人们认为，在该宗中，200户是能够实现的最高样本量。
基于多个考虑因素，包括设计效应、可用的预算以及每个团队完成一个集群所需的时间，BMIS中每个集群选定的家庭平均数量确定为20户。通过将家庭总数除以每个集群的样本家庭数量，计算得出每个宗（除加萨外）需要选择40个样本集群。在加萨，需要选择10个。
除了加萨外，将总样本量平均分配给20个宗。因此，除了加萨（10个）外，每个宗分配了40个集群，最终计算出的样本量为15,400户：40个集群 * 19个宗 * 每个集群20个样本家庭，加萨10个集群 * 1个宗 * 每个集群20个样本家庭。在每个宗中，集群（一级抽样单位）被分配到城市和农村领域，比例与该宗城市和农村人口规模相对应。下表显示了集群的分配情况。
抽样程序在《不丹多指标集群调查2010 - 报告》的第207-212页有更详细的描述。
数据收集方式
---------------------------
面对面 [f2f]
研究工具
---------------------------
通用MICS的问卷是基于MICS4模型问卷的结构化问卷，并进行了一些修改和补充。家庭问卷由居住在家庭中的有知识的人士填写。家庭问卷包括家庭登记表、教育、水和卫生、家庭特征、儿童劳动、残疾和洗手。
除了家庭问卷外，还为每个家庭中的15-49岁的妇女和5岁以下儿童进行了问卷调查。5岁以下儿童的问卷由所有居住在家庭中的5岁以下儿童的母亲或照顾者填写。
妇女问卷包括妇女背景、儿童死亡率、对最后一次生育的愿望、孕产妇和新生儿健康、避孕、未满足的需求、婚姻/联合、对家庭暴力的态度和经历、性行为、HIV/AIDS和孕产妇死亡率。
儿童问卷包括儿童年龄、出生登记、早期儿童发展、母乳喂养、疾病护理和人体测量学。
数据清理操作
---------------------------
使用CSPro软件在25台微型计算机上录入数据，由25名操作员和3名监督员进行录入。为了确保质量控制，所有问卷都进行了双录入，并进行了内部一致性检查。在整个过程中使用了全球MICS4计划下开发的程序和标准程序，并对其进行了调整以适应不丹问卷。数据录入始于数据收集开始后一个月，并于2010年9月完成。使用预测分析软件（PASW）、SPSS软件的第18版进行分析，并使用了联合国儿童基金会为此目的开发的模型语法和表格计划。
响应率
---------------------------
在选定的15,400户样本中，14,917户有人居住。其中，14,676户成功接受了访谈，家庭响应率为98.4％。在受访的家庭中，确定了16,823名符合条件的妇女（年龄在15-49岁之间）。其中14,018人成功接受了访谈，响应率为83.3％。家庭访谈确定了6,457名5岁以下儿童。其中6,297名儿童的问卷已填写，响应率为97.5％。
抽样误差估计
---------------------------
抽样误差是衡量所有可能样本估计之间差异的一种度量。差异的程度并不完全清楚，但可以从调查数据中统计地估计。
本附录中为每个选定的指标提供了以下抽样误差度量：
• 标准误差（se）：抽样误差通常以特定指标的（平均值、比例等）标准误差来衡量。标准误差是估计值方差的平方根。使用泰勒线性化方法来估计标准误差。
• 变异系数（se/r）是标准误差与指标值的比率，是相对抽样误差的度量。
• 设计效应（deff）是使用调查中采用的抽样方法下指标的实际情况方差与在简单随机抽样假设下计算的方差之比。设计效应（deft）的平方根用于表明样本设计在精度方面的效率。deft值为1.0表示样本设计与简单随机样本一样有效，而deft值大于1.0表示由于使用了更复杂的样本设计而导致的标准误差增加。
• 置信区间是计算出来的，以显示真实值在给定置信水平下可以合理假设落在其中的区间。对于从调查中计算出的任何给定统计量，该统计量的值将在95％的所有可能样本（大小和设计相同）的正负两倍标准误差（r + 2.se或r - 2.se）范围内。
在计算MICS数据的抽样误差时，使用了SPSS版本18的复杂样本模块。结果如下表所示。除了上述抽样误差度量外，表格还包括每个指标的加权计数和无权计数。
抽样误差针对主要指标的国家级、地区级和城市与农村地区进行了计算。八个选定的指标基于家庭成员，十八个基于妇女，十二个基于5岁以下儿童。此处呈现的所有指标均以比例的形式。
数据评估
---------------------------
一系列数据质量表可用于审查数据质量，包括以下内容：
- 家庭人口年龄分布
- 符合条件和接受访谈的妇女年龄分布
- 家庭中5岁以下儿童和5岁以下儿童问卷的年龄分布
- 妇女完成率按家庭社会经济特征
- 5岁以下儿童问卷完成率按家庭社会经济特征
- 报告的完整性
- 人体测量学指标信息的完整性
- 人体测量学测量的堆叠
- 观察洗手地点
- 观察妇女健康卡
- 观察5岁以下儿童出生证明
- 家庭中母亲的存在和接受5岁以下儿童问卷访谈的人
- 单个年龄段的学校出席率
- 所有出生并存活下来的儿童的出生性别比

这些数据质量表的结果在“不丹多指标集群调查2010 - 最终报告”的附录D中展示。