Multiple Indicator Cluster Survey 2000 - Lesotho

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

This Draft Preliminary Report presents the initial results of the 2000 Lesotho End-Decade Multiple Indicator Cluster Survey (EMICS). These results were derived from a nationally representative survey of households, women, and children. The main objectives of the survey were:

- to provide up-to-date information for assessing the situation of children and women in Lesotho at the end of the decade, and;
- to furnish data needed for monitoring progress toward goals established at the World Summit for Children and as a basis for future action.

It is organised in four main sections. Chapter One is a documentation of the background to the 2000 Lesotho EMICS, and its objectives. The technical details of the survey, including sampling procedures, data collection and analysis are reported in Chapter Two. Chapter Three presents an evaluation of the quality of the data collected during the survey and used to prepare the analysis presented in Chapter Four. The Appendices attach the questionnaire used in the survey, some key documentation and lists key personnel and organisations involved in the 2000 Lesotho EMICS. It is expected that this Draft Preliminary Report will generate discussion on the findings in respect of the health, education, and child labour situation in Lesotho amongst government agencies, non-governmental organisations (NGOs), multilateral donors, the press and the public.

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

The Lesotho MICS survey was a nationally representative survey of households, women, and children.

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

households, Women, Children.

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

Sample survey data [ssd]

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

The Bureau of Statistics provided the information that was utilised in constructing the 2000 Lesotho MICS, including the Rural Master Sampling Frame, the 1996 Population Census Frame, Urban Population Sampling Units File. Information availed through these documents included Enumeration Area (EA) Numbers, Number of Villages in the EAs, and the Number of Households. Table 2.0 shows the characteristics of this constructed sampling frame. The 2000 Lesotho EMICS covered the whole country. Thus, the sample was designed to provide estimates of health indicators at four levels.The national level, the urban and rural levels, the regional level the district level. The largest sample size was regarded as the required sample size that would provide adequate information on all the indicators. It would have been ideal to maintain the distinction in the four ecological strata, however for practical reasons this was not possible. Senqu River Valley and the Mountains were combined, because having them reported separately, would have required a much larger sample to detect differences between the two. Two stage cluster sampling was applied in selecting the sample for the 2000 Lesotho EMICS exercise, using an updated sampling frame from the 1996 population census (see Table 2.0). A number of variables indicators and assumptions were taken into consideration when determining the sample size for the 2000 Lesotho EMICS. These included:
- The number of households by district, both urban and rural
- Ecological considerations were employed in the rural areas
- Measles was used as a key indicator
- Design effect of 2 for most variables
- Level of estimation. In addition to national estimates, district estimates for comparison were required.
- The error margins were approximately 10 percent for district level data and five percent for national level data

Thus, the ecological strata were used for the ten districts, which were the main domains. Enumeration areas were the first stages of selecting sampling units on a systematic basis. Selection of primary sampling units was from each of the four ecological zones in the rural and urban areas. The second stage involved a systematic selection of twenty (20) households to cover the entire selected village of the selected enumeration areas. Consequently, 380 PSUs were identified, from which in each PSU twenty households were systematically selected, to give an expected sample of 7,600 households.
The systematic selection of the households involved a listing procedure, where all the households in the selected village were listed. A systematic sample was to be selected and interviewed. To achieve this two- (2) enumerators were selected and assigned to each village to undertake the listing of households and were provided with a map or a sketch showing the boundaries of the areas under consideration.
Once the listing was accomplished the supervisor collected the completed household listing forms (Form 1 - see Appendix C) and together with the enumerator systematically drew a sample of twenty (20) households from each selected village.

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

Face-to-face [f2f]

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

The questionnaire applied in the 2000 Lesotho EMICS comprised a household questionnaire, questionnaires were administered in each household for women aged 15 – 49 and children under age five. The questionnaires were based on the MICS model questionnaire with additional modules. In all the following modules and panels constituted the 2000 Lesotho EMICS questionnaire5. The questionnaire was translated into the Sesotho Language. A pretest of the questionnaire was performed on the 21st of January 2000. It involved two teams of three each supervised by the National Co-ordinator, EMICS and the MICS Consultant. The pretest was done in Ha Foso Village of Berea district, which was selected for its accessibility to Maseru. Questionnaires were pretested for questionnaire administration time, key questions were tested for reliability and consistency and last but not least the logistics for successful questionnaire administration were tested. Based on the findings of this pretest, modifications to the questionnaire were made in respect of the wording and the translation as well as instructions to the enumerators aimed at ensuring proper questionnaire administration. Selection of the interview households was randomly done.

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

Overall the household response rate is about 99.7 percent, the eligible women’s response rate is 93.6 percent, and that of under five children is 98.0 percent. This are higher than the 90 percent response rates normally expected from surveys similar to the 2000 Lesotho EMICS. In respect of the households, the planned sample was to select and interview 7,600 households. Based on this planned figure the response rate would be a slightly lower 97.4 percent, but still higher than the base 90 percent for similar surveys. One can therefore conclude that the 2000 Lesotho EMICS had a good response rate and its results would be, all things considered, reliable.

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

Sampling errors on the other hand can be evaluated statistically. Consider the fact that the 2000 Kingdom of Lesotho MICS Sample is only one of the many possible samples that could have been chosen using the same sample design and expected size. Sampling errors in this regard are a measure of the variability between all possible samples. These sampling errors are usually measured in terms of the standard errors for a particular statistic (mean, percentage, etc.), which is the square root of the variance. The standard error can be used to determine the confidence intervals within which the true value for the population can be reasonably assumed to fall.

摘要
---------------------------
本草案初步报告展示了2000年莱索托跨世纪多指标聚类调查（EMICS）的初步结果。这些结果源自对全国范围内的家庭、妇女和儿童的代表性调查。调查的主要目标包括：
- 提供最新信息，以评估莱索托在世纪末儿童和妇女的状况，以及；
- 提供数据，用于监测实现世界儿童峰会所确立目标的过程，并为未来的行动奠定基础。
报告分为四个主要部分。第一章记录了2000年莱索托EMICS的背景及其目标。第二章报告了调查的技术细节，包括抽样程序、数据收集和分析。第三章对调查收集的数据质量以及用于准备第四章分析的数据进行了评估。附录包含了调查中使用的问卷、一些关键文件，并列出了参与2000年莱索托EMICS的关键人员和组织。预计本草案初步报告将引发政府机构、非政府组织（NGO）、多边捐助者、新闻界和公众就莱索托在健康、教育和童工状况方面的发现进行讨论。
地理覆盖范围
---------------------------
莱索托MICS调查是对家庭、妇女和儿童的全国代表性调查。
分析单元
---------------------------
家庭、妇女、儿童。
数据类型
---------------------------
样本调查数据 [ssd]
抽样程序
---------------------------
统计局提供了构建2000年莱索托MICS所需的信息，包括农村抽样框架、1996年人口普查框架和城市人口抽样单位文件。这些文件提供的信息包括：
- 编号区域（EA）编号
- EA中的村庄数量
- 家庭数量。
表2.0显示了此构建抽样框架的特征。2000年莱索托EMICS涵盖了整个国家。因此，样本旨在提供四个级别的健康指标估计值：国家层面、城市和农村层面、地区层面和区级层面。最大的样本量被认为是提供所有指标充足信息的所需样本量。理想情况下，应保持四个生态层级的区分，但出于实际考虑，这并不可能。森昆河谷和山区被合并，因为如果单独报告，将需要更大的样本才能检测到两者之间的差异。在2000年莱索托EMICS活动中应用了两阶段聚类抽样，使用的是1996年人口普查的更新后的抽样框架（见表2.0）。在确定2000年莱索托EMICS的样本量时，考虑了以下一些变量、指标和假设：
- 各个地区的家庭数量，包括城市和农村
- 在农村地区应用了生态考虑因素
- 麻疹被用作关键指标
- 大多数变量的设计效应为2
- 估计水平。除了国家估计值外，还需要区级估计值进行比较。
- 区级数据的误差范围约为10%，国家数据为5%。
因此，生态层被用于十个区，这些是主要领域。编号区域是选择抽样单位的第一阶段。从农村和城市地区的四个生态区域中选取了初级抽样单位。第二阶段涉及在所选编号区域中选择二十（20）个家庭，以覆盖所选村庄的整个村庄。因此，确定了380个PSU，每个PSU中系统地选择了二十个家庭，以提供一个预期的样本量为7,600个家庭。
家庭的选择涉及一个列表程序，其中列出了所选村庄中的所有家庭。应选择并访谈的系统样本。为此，选出了两名（2）调查员并分配给每个村庄，以执行家庭列表，并提供了显示所考虑区域边界的地图或草图。
一旦完成列表，监督员收集了完成的家庭列表表（附录C中的表格1）和调查员，系统地从每个所选村庄中抽取了二十（20）个家庭。
数据收集方式
---------------------------
面对面 [f2f]
研究工具
---------------------------
2000年莱索托EMICS中使用的问卷包括家庭问卷，对每个家庭中的15-49岁妇女和5岁以下儿童进行了问卷。问卷基于MICS模型问卷，并增加了额外的模块。以下所有模块和面板构成了2000年莱索托EMICS问卷。问卷被翻译成塞索托语。2000年1月21日对问卷进行了预测试。涉及两个由EMICS国家协调员和MICS顾问监督的每组三人的团队。预测试在贝雷区的哈福索村进行，该村因其对马萨鲁的可及性而被选中。对问卷进行了预测试，以评估问卷的行政时间，对关键问题进行了可靠性测试，最后但并非最不重要的是，对成功进行问卷的物流进行了测试。根据预测试的结果，对问卷的文字、翻译以及对调查员的指导进行了修改，以确保适当的问卷管理。访谈家庭的选择是随机进行的。
应答率
---------------------------
总体而言，家庭应答率约为99.7%，合格妇女的应答率为93.6%，5岁以下儿童的应答率为98.0%。这些应答率高于类似2000年莱索托EMICS调查通常期望的90%应答率。就家庭而言，计划样本是为了选择和访谈7,600个家庭。根据这个计划数字，应答率略低为97.4%，但仍然高于类似调查的基础90%。
抽样误差估计
---------------------------
另一方面，抽样误差可以从统计上进行评估。考虑一下，2000年莱索托王国MICS样本只是使用相同样本设计和预期规模可能选择的许多样本之一。这方面的抽样误差是衡量所有可能样本之间差异的度量。这些抽样误差通常以特定统计量（均值、百分比等）的标准误差来衡量，这是方差的平方根。标准误差可用于确定置信区间，在此区间内可以合理地假设真实值对于总体。