Demographic and Health Survey 1986 - Liberia

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

The Liberia Demographic and Health Survey (LDHS) was conducted as part of the worldwide Demographic and Health Surveys (DHS) program, in which surveys are being carried out in countries in Africa, Asia, Latin America, and the Middle East. Liberia was the second country to conduct a DHS and the first country in Africa to do so. THe LDHS was a national-level survey conducted from February to July 1986, covering a sample of 5,239 women aged 15 to 49.

The major objective of the LDHS was to provide data on fertility, family planning and maternal and child health to planners and policymakers in Liberia for use in designing and evaluating programs. Although a fair amount of demographic data was available from censuses and surveys, almost no information existed concerning family planning, health, or the determinants of fertility, and the data that did exist were drawn from small-scale, sub-national studies. Thus, there was a need for data to make informed policy choices for family planning and health projects.

A more specific objective was to provide baseline data for the Southeast Region Primary Health Care Project. In order to effectively plan strategies and to eventually evaluate the progress of the project in meeting its goals, there was need for data to indicate the health situation in the two target counties prior to the implementation of the project. Many of the desired topics, such as immunizations, family planning use, and prenatal care, were already incorporated into the model DHS questionnaire; nevertheless, the LDHS was able to better accommodate the needs of this project by adding several questions and by oversampling women living in Sinoe and Grand Gedeh Counties.

Another important goal of the LDHS was to enhance tile skills of those participating in the project for conducting high-quality surveys in the future. Finally, the contribution of Liberian data to an expanding international dataset was also an objective of the LDHS.

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

National

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

- Households
- Children age 0-5
- Women age 15 to 49
- Men

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

Sample survey data

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

The sample for the Liberia Demographic and Health Survey was based on the sampling frame of about 4,500 censal enumeration areas (EAs) that were created for the 1984 Population Census. It was decided to eliminate very remote EAs prior to selecting the sample. The definition of remoteness used was "any EA in which the largest village was estimated to be more than 3-4 hours' walk from a road." According to the 1984 census, the excluded areas represent less than 3 percent of the total number of households in the country. Since the major analytic objective of the LDHS was to adequately estimate basic demographic and health indicators including fertility, mortality, and contraceptive prevalence for the whole country and the two sub-universes (Since and Grand Gedeh Counties), it was decided to oversample these two counties. Consequently, three explicit sub-universes of EAs were created: (1) Since County, (2) Grand Gedeh County, and (3) the rest of the country.

The design provided a self-weighted sample within each sub-universe, but, because of the oversampling in Sinoe and Grand Gedeh Counties, the sample is not self-weighting at the national level. Eligible respondents for the survey were women aged 15-49 years who were present the night before the interview in any of the households included in the sample selected for the LDHS.

The total sample size was expected to be about 6,000 women aged 15-49 with a target by sub-universe of 1,000 each in Sinoe and Grand Gedeh Counties and 4,000 in the rest of the country. It was decided that a sample of approximately 5,500 households selected through a two-stage procedure would be appropriate to reach those objectives. Sampling was carried out independently in each sub-universe. In the rest of the country sub-universe, counties were arranged for selection in serpentine order from the northwest (Cape Mount County) to the southeast (Maryland County). In the first stage EAs were selected systematically with probability proportional to size (size = number of households in 1984). Twenty-four EAs were selected in each of Sinoe and Grand Gedeh Counties and 108 EAs in the rest of the country.

See full sample procedure in the survey final report.

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

Face-to-face

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

The Liberia Demographic and Health Survey (LDHS) utilized two questionnaires: One to list members of the selected households (Household Questionnaire) and the other to record information from all women aged 15-49 who were present in the selected households the night before the interview (Individual Questionnaire).

Both questionnaires were produced in Liberian English and were pretested in September 1985. The Individual Questionnaire was an early version of the DHS model questionnaire. It covered three main topics: (1) fertility, including a birth history and questions concerning desires for future childbearing, (2) family planning knowledge and use, and (3) family health, including prevalence of childhood diseases, immunizations for children under age five, and breasffeeding and weaning practices.

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

Data from the questionnaires were entered onto microcomputers at the Bureau of Statistics office in Monrovia. The data were then subjected to extensive checks for consistency and accuracy.

Errors detected during this operation were resolved either by referring to the original questionnaire, or, in some cases, by logical inference from other information given in the record. Finally, dates were imputed for the small number of cases where complete dates of important events were not given.

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

Out of the total of 6,1306 households selected, 14.5 percent were found not to be valid households in the field, either because the dwelling had been vacated or destroyed, or the household could not be located or did not exist. Of the 5,609 households that were found to exist, 90 percent were successfully interviewed. In the households that were interviewed, a total of 5,340 women were identified as being eligible for individual interview (that is, they were aged 15-49 and had spent the night before the interview in the selected household). This represents an average of slightly over one eligible woman per household.

The response rate for eligible women was 98 percent. The main reason for nonresponse was the absence of the woman. Similar data are presented by sample subuniverse.

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

The results from sample surveys are affected by two types of errors: (1) nonsampling error and (2) sampling error. Nonsampling error is due to mistakes made in carrying out field activities, such as failure to locate and interview the correct household, errors in the way questions are asked, misunderstanding of the questions on the part of either the interviewer or the respondent, data entry errors, etc. Although efforts were made during the design and implementation of the Liberia Demographic and Health Survey to minimize this type of error, nonsampling errors are impossible to avoid and difficult to evaluate statistically.

The sample of women selected in the LDHS is only one of many samples of the same size that could have been selected from the same population, using the same design. Each one would have yielded results that differed somewhat from the actual sample selected. The variability observed between all possible samples constitutes sampling error, which, although it is not known exactly, can be estimated from the survey results. Sampling error is usually measured in terms of the "standard error" of a particular statistic (mean, percentage, etc.), which is the square root of the variance of the statistic across all possible samples of equal size and design.

The standard error can be used to calculate confidence intervals within which one can be reasonably assured the true value of the variable for the whole population falls. For example, for any given statistic calculated from a sample survey, the value of that same statistic as measured in 95 percent of all possible samples of identical size and design will fall within a range of plus or minus two times the standard error of that statistic.

If the sample of women had been selected as a simple random sample, it would have been possible to use straightforward formulas for calculating sampling errors. However, the LDHS sample design depended on stratification, stages, and clusters and consequently, it was necessary to utilize more complex formulas. The computer package CLUSTERS was used to assist in computing the sampling errors with the proper statistical methodology.

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

Information on the completeness of date reporting is of interest in assessing data quality. With regard to dates of birth of individual women, 42 percent of respondents reported both a month and year of birth, 21 percent gave a year of birth in addition to current age, and 37 percent gave only their ages. With regard to children's dates of birth in the birth history, 85 percent of births had both month and year reported, 12 percent had year and age reported, 1 percent had only age reported, and 2 percent had no date information.

摘要
---------------------------
利比里亚人口与健康调查（LDHS）作为全球人口与健康调查（DHS）项目的一部分而进行，该项目在非洲、亚洲、拉丁美洲和中东的国家进行调查研究。利比里亚是第二个实施DHS的国家，也是非洲第一个实施的国家。LDHS于1986年2月至7月在全国范围内进行，调查了5,239名15至49岁的女性。
LDHS的主要目标是向利比里亚的规划者和政策制定者提供关于生育、家庭规划和母婴健康的数据，以便用于设计和发展评估项目。尽管从人口普查和调查中可以获得相当数量的统计数据，但关于家庭规划、健康或生育决定因素的信息几乎不存在，而现有数据大多来源于小规模的次国家级研究。因此，需要数据来为家庭规划和健康项目做出明智的政策选择。
更具体的目标是为东南地区初级卫生保健项目提供基线数据。为了有效地制定策略并最终评估项目实现目标的情况，需要数据来表明项目实施前两个目标县的健康状况。许多期望探讨的主题，如疫苗接种、家庭规划使用和产前护理，已纳入模型DHS问卷；然而，LDHS通过增加一些问题并增加西诺县和格兰德吉德县妇女的样本量，更好地满足了该项目的需求。
LDHS的另一个重要目标是提高参与项目人员在今后开展高质量调查的技能。最后，将利比里亚数据贡献给不断扩大的国际数据集也是LDHS的目标。
地理覆盖范围
---------------------------
全国
分析单位
---------------------------
- 家庭
- 0-5岁儿童
- 15至49岁女性
- 男性
数据类型
---------------------------
样本调查数据
抽样程序
---------------------------
利比里亚人口与健康调查的样本基于为1984年人口普查创建的大约4,500个普查区（EA）的抽样框架。在选定样本之前，决定消除非常偏远的EA。使用的偏远定义是“任何最大村庄距离道路估计超过3-4小时步行时间的EA”。根据1984年的人口普查，被排除的地区占全国总户数的不到3%。由于LDHS的主要分析目标是充分估计包括生育、死亡率和避孕普及率在内的基本人口和健康指标，包括整个国家和两个子集（西诺县和格兰德吉德县），因此决定对这两个县进行过度抽样。因此，创建了三个显式的EA子集：（1）西诺县，（2）格兰德吉德县，（3）该国其余地区。
设计在每个子集中提供了一个自加权样本，但由于西诺县和格兰德吉德县的过度抽样，样本在国家层面上不是自加权的。合格的受访者是那些在LDHS所选样本家庭中居住并在采访前一夜出现的15-49岁女性。
预期总样本量约为6,000名15-49岁的女性，目标是在西诺县和格兰德吉德县各有1,000名，在其他地区有4,000名。决定通过两阶段程序选择大约5,500户家庭以实现这些目标。在每个子集中独立进行抽样。在其他地区子集中，从西北部（卡普蒙县）到东南部（马里兰州）按蛇形顺序排列县进行选择。在第一阶段，EA按规模成比例的概率系统选择（规模=1984年家庭数量）。西诺县和格兰德吉德县各选择了24个EA，其他国家选择了108个EA。
在调查最终报告中查看完整的抽样程序。
数据收集方式
---------------------------
面对面
研究工具
---------------------------
利比里亚人口与健康调查（LDHS）使用了两个问卷：一个用于列出所选家庭的成员（家庭问卷），另一个用于记录在所选家庭中采访前一晚出现的所有15-49岁女性的信息（个人问卷）。
两个问卷都用利比里亚英语制作，并于1985年9月进行了预测试。个人问卷是DHS模型问卷的早期版本。它涵盖了三个主要主题：（1）生育，包括出生史和有关未来生育意愿的问题，（2）家庭规划知识和使用，（3）家庭健康，包括儿童疾病的发生率、五岁以下儿童的疫苗接种以及母乳喂养和断奶实践。
数据清洗操作
---------------------------
问卷数据在蒙罗维亚统计局办公室的微计算机上录入。然后对这些数据进行广泛的检查，以确保一致性和准确性。
在此次操作中发现的错误要么通过查阅原始问卷来解决，要么在某些情况下，通过从记录中提供的其他信息进行逻辑推断来解决。最后，对那些重要事件的完整日期未给出的小部分案例进行了日期的推定。
响应率
---------------------------
在总共6,1306户选定家庭中，有14.5%的家庭在实地中被发现不是有效家庭，要么是因为住宅已被空置或摧毁，要么是因为家庭无法找到或不存在。在发现的5,609户家庭中，90%的家庭被成功采访。在采访的家庭中，共确定了5,340名符合条件的女性接受个人采访（即她们年龄在15-49岁之间，并在所选家庭中度过采访前一晚）。这代表平均每户略超过一名符合条件的女性。
符合条件的女性的响应率为98%。非响应的主要原因是没有女性在场。按样本子集提供了类似的数据。
抽样误差估计
---------------------------
样本调查的结果受到两种类型的误差的影响：（1）非抽样误差和（2）抽样误差。非抽样误差是由于在现场活动中犯的错误造成的，例如未能找到和采访正确的家庭，提问方式错误，采访者或受访者对问题的误解，数据录入错误等。尽管在设计和发展利比里亚人口与健康调查时努力减少此类误差，但非抽样误差是不可避免的，并且难以从统计上进行评估。
LDHS中选定的女性样本只是从同一人口中用相同设计选出的许多相同大小的样本之一。每一个都会产生与实际选定的样本略有不同的结果。所有可能样本之间的可变性构成了抽样误差，尽管它不完全为人所知，但可以从调查结果中估计。抽样误差通常用特定统计量（均值、百分比等）的“标准误差”来衡量，它是所有可能样本的方差平方根，这些样本的大小和设计相同。
标准误差可用于计算置信区间，人们可以合理地保证变量在总体中的真实值落在该区间内。例如，对于从样本调查中计算出的任何给定的统计量，该统计量在所有可能样本的95%中测量的值将落在该统计量的标准误差的两倍范围内。
如果女性的样本是作为简单随机样本选择的，那么就可以使用简单的公式来计算抽样误差。然而，LDHS的样本设计依赖于分层、阶段和聚类，因此有必要使用更复杂的公式。使用计算机软件包CLUSTERS来协助使用正确的统计方法计算抽样误差。
数据评估
---------------------------
关于日期报告完整性的信息在评估数据质量时很有趣。就个人女性的出生日期而言，42%的受访者报告了出生月份和年份，21%的人除了当前年龄外还报告了出生年份，37%的人只报告了年龄。就出生史中儿童的出生日期而言，85%的出生都报告了月份和年份，12%的出生报告了年份和年龄，1%的出生只报告了年龄，2%的出生没有日期信息。