Child Labor Survey 2001 - Cambodia

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

The Royal Government of Cambodia has placed the issue of child labour 2001 high on its agenda, and in recent years has made a large number of interventions, in cooperation with international, regional organizations, NGOs, and other partners. For Cambodia, like the rest of the world, there are many issues that need to be addressed in order to ensure the effective abolition of child labour. Eliminating child labour in Cambodia is one of the most urgent challenges of our time.

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

National Phnom Penh, Other Urban and Other Rural area Urban/Rural areas

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

- Household
- Individual
- Children aged 5-17 years

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

The survey covered all household members and children aged 5 to 17 years residing in the household.

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

Sample survey data [ssd]

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

The sample for CCLS 2001 was a stratified sample selected in two stages. At the first stage, the villages (primary sampling units or PSU's) were selected from the list of villages for every stratum within the domains listed in order of: province/city, district, commune and village. The method of circular systematic sampling with the probability of inclusion of a village proportional to its size (CSS-PPS) was used to select the villages. In addition to the code for the above, the sampling frame contained identification particulars: the name of the villages as well as the number of households in the village as known at that time. The number of households in the village identified in the sampling frame was taken as its size. The actual number of households in selected villages was obtained later by direct listing. In this method, the actual number of households was generally different from the number recorded in the sampling frame. To avoid ambiguity, the number of households in a village as recorded in the sampling frame will always be referred to as its size. In other words, the number of households in the village was used as the measure of size. Sample village selection was done through the use of a computer program.

At the second stage of selection, for each sample village, or PSU (except for large villages), a field listing operation was undertaken. Large villages comprising about 210 households, based on the current household estimates by the village leaders, were segmented first. A segment was then chosen randomly and a complete listing of households was prepared. This entailed carrying out a complete canvass of the PSU in order to make a current and complete listing of households contained within. The procedure involved creating a sketch map of the PSU where physical boundaries in the village and the location of each household were sketched. Canvassing entailed a systematic covering of the entire village following a prescribed path of travel in order to make sure that all housing units in which the households reside were accounted for. After the listing operation was completed, a fixed sample size of 20 households was selected in each PSU. The selection was carried out using a circular systematic random sampling with a random start (CSS).

Detailed discussion is provided in Sampling.pdf document.

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

Face-to-face [f2f]

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

The Cambodia Child Labour survey 2001 used three forms:
Form 1: Questionnaire for Households
Form 2: Questionnaire for Parents/Guardians of Children Ages 5 to 17 in the Household
Form 3: Questionnaire for Children Ages 5 to 17 in the Household

The household questionnaire (Form 1) included demographic characteristics of all household members, migration status of the household (in the last five years), migration status of children 5 to 17 years of age, housing conditions, household income and expenditures, usual economic activity of household members 5 years old and over during the last 12 months, current economic activity of household members 5 years and over during the past 7 days, earnings and hours of work of household members 5 years old and over who worked (according to the main activity during the past 7 days and other activities during the past 7 days).

In the questionnaire about children aged 5 to 17 in the households (Form 2), queries all children 5 to 17 years old were addressed to parents, guardians, or responsible proxies in the household where the child usually resides. This form contained questions on housekeeping activities/household chores of children 5 to 17 years old; children 5 to 17 years old who were idle/did not do anything during the past 7 days; health and safety of children 5 to 17 years old who had worked at any time; place of work/employer of children 5 to 17 years of age currently working; children 5 to 17 years working as employees for someone else for payment in cash or in kind or without any payment; perception of parents/guardians or other relatives with whom the working child usually resides.

In the questionnaire for children (Form 3), the questions were addressed to children 5 to 17 years of age: hours actually worked, economic activities, age started to work for the first time, currently or usually worked, health and safety of working children, working conditions and perceptions of working children.

The statistical information on hazardous child labour was collected through an establishment survey. Selection of establishments was based on the results of the household-based child labour survey, and establishments found to employ child labour were identified for a follow-up survey by interviewers with the employers and the children identified as child workers.

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

Processing of the Cambodia Child Labour Survey 2001, from data entry to tabulations, was done with the use of the Integrated Microcomputer Processing System (IMPS) of the US Bureau of Census. 20 Machine Encoders (including 2 Supervisors) were trained from 23/05/01 to 25/05/01 before the start of machine encoding activities.

The data dictionary and record design of the questionnaire was developed using IMPS, and the data entry screen was designed to enter information from the questionnaires.

To confirm the legitimacy of the entries, the data in the questionnaires was reentered. The data files created through this system are in ASCII format and the variables in each data file can be identified using the data dictionary that is based on IMPS. Database files in Foxpro/Foxbase/dBase format was also created.

To check the correctness, completeness, range validity and consistency between items of the data, an editing program was developed. This software was created to check and validate the data files, and to produce error printouts for modification. The cleaned database was compared with other sources before it was used for further processing.

Data editing took place at the following stages of data processing:
1. On the field: Manual editing on the done by supervisors.
2. Questionnaire reception at NIS: Manual editing and coding done by 15 manual editing personnel at NIS.
3. During data entry: Re-entry of data in the questionnaires to confirm legitimacy of the entries.
4. Correctness, completeness, range validity and consistency checking: An editing program was developed to check and validate data files and produce error printouts from modification.
5. Verification of database with other sources
6. Data quality tabulation: With the preliminary tables generated, data validation work was carried out and range checks built were reviewed.

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

The need to adjust the weights for non-response did not arise as completed questionnaires from all sampled villages and households were retrieved achieving a 100% response rate.

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

The estimation of the magnitude of sampling error in the survey data was computed. Since most of the estimates from the survey are in the form of weighted ratios, variances for ratio estimates were thus presented. Appendix A of the CCLS 2001 report discusses in detail the sampling error estimation.

摘要
---------------------------
柬埔寨王国政府于2001年将儿童劳动问题提上议程，并在近年与各国、地区组织、非政府组织和其它合作伙伴合作，实施了大量干预措施。对于柬埔寨而言，如同世界其它地区，为确保有效废除儿童劳动，仍有许多问题亟待解决。在柬埔寨废除儿童劳动是我们时代面临的最紧迫挑战之一。
地理覆盖范围
---------------------------
全国：金边，其他城市和其它农村地区，城乡地区
分析单位
---------------------------
- 家庭
- 个人
- 5至17岁儿童
总体
---------------------------
本调查覆盖了居住在家庭中的所有家庭成员和5至17岁的儿童。
数据类型
---------------------------
样本调查数据 [ssd]
抽样程序
---------------------------
CCLS 2001的样本为两阶段分层抽样。在第一阶段，从每个层级的村庄列表中（一级抽样单位或PSU）选择村庄，层级按以下顺序排列：省份/城市、区、乡和村庄。使用与村庄规模成比例的包含概率的循环系统抽样（CSS-PPS）方法选择村庄。除了上述代码外，抽样框架还包含识别细节：村庄名称以及当时所知的村庄家庭数量。抽样框架中确定的村庄家庭数量被视为其规模。后来通过直接清点获得了所选村庄的实际家庭数量。在此方法中，实际家庭数量通常与抽样框架中记录的数量不同。为了避免歧义，抽样框架中记录的村庄家庭数量将始终被称为其规模。换句话说，村庄的家庭数量被用作规模的衡量标准。通过计算机程序进行样本村庄的选择。
在选择的第二阶段，对每个样本村庄或PSU（除大村庄外），进行了现场清点操作。根据村长目前的家庭估计，拥有大约210户家庭的大村庄首先被分割。然后随机选择一个分割区，并编制了完整的家庭清单。这涉及对PSU进行全面清点，以便编制包含在内的家庭清单。该程序包括绘制PSU的草图地图，其中绘制了村庄的物理边界和每个家庭的地理位置。清点涉及遵循规定的旅行路线系统地覆盖整个村庄，以确保计入所有居住在住房单位中的家庭。清点操作完成后，每个PSU选择了固定样本量的20户家庭。选择是通过使用循环系统抽样（CSS）和随机起点进行的。详细讨论见Sampling.pdf文件。
数据收集方式
---------------------------
面对面 [f2f]
研究工具
---------------------------
柬埔寨儿童劳动调查2001年使用了三种表格：
表格1：家庭问卷
表格2：针对家庭中5至17岁儿童的父母/监护人的问卷
表格3：针对家庭中5至17岁儿童的问卷
家庭问卷（表格1）包括所有家庭成员的人口统计特征，家庭在过去五年内的迁移状况，5至17岁儿童的迁移状况，住房条件，家庭收入和支出，5岁及以上家庭成员在过去12个月内的通常经济活动，5岁及以上家庭成员在过去7天内的当前经济活动，5岁及以上工作家庭成员的工资和工作小时数（根据过去7天的主要活动和其他活动）。
在关于家庭中5至17岁儿童的问卷（表格2）中，对通常居住在家庭中的5至17岁儿童的家长、监护人或负责的代理人进行了询问。本表格包含有关5至17岁儿童的家务活动/家庭琐事；过去7天无所事事/什么都没做的5至17岁儿童；任何时间工作的5至17岁儿童的卫生和安全；目前工作的5至17岁儿童的 workplaces/雇主；5至17岁作为雇员为他人工作并获得现金或实物或无报酬的儿童；居住在通常与工作儿童同住的父母/监护人或其他亲属的看法。
在针对儿童的问卷（表格3）中，问题针对5至17岁的儿童：实际工作时间，经济活动，第一次开始工作的时间，目前或通常工作，工作儿童的卫生和安全，工作条件和劳动儿童的看法。
通过企业调查收集有关危险儿童劳动的统计数据。企业的选择基于基于家庭的儿童劳动调查结果，发现雇佣儿童劳动的企业被识别出来，以便通过访谈雇主和被认定为儿童工人的儿童进行后续调查。
数据清洗操作
---------------------------
柬埔寨儿童劳动调查2001年的数据处理，从数据录入到制表，均使用美国人口普查局的集成微机处理系统（IMPS）进行。在开始机器编码活动之前，从2001年5月23日至25日对20名机器编码员（包括2名主管）进行了培训。
使用IMPS开发了数据字典和问卷记录设计，并设计了数据录入屏幕，以从问卷中输入信息。
为了确认输入的合法性，重新输入了问卷中的数据。通过此系统创建的数据文件为ASCII格式，每个数据文件中的变量可以使用基于IMPS的数据字典进行识别。还创建了Foxpro/Foxbase/dBase格式的数据库文件。
为了检查正确性，完整性，范围有效性和项目之间的一致性，开发了编辑程序。该软件被创建用于检查和验证数据文件，并从修改中生成错误打印输出。在进一步处理之前，清理后的数据库与其它来源进行了比较。
数据编辑发生在数据处理的以下阶段：
1. 在现场：由主管进行的手动编辑。
2. 在国家统计局接收问卷时：国家统计局的15名手动编辑人员进行的手动编辑和编码。
3. 在数据录入期间：为了确认输入的合法性，重新输入问卷中的数据。
4. 正确性，完整性，范围有效性和一致性检查：开发了编辑程序来检查和验证数据文件，并从修改中生成错误打印输出。
5. 与其它来源验证数据库
6. 数据质量制表：通过生成的初步表格，进行了数据验证工作，并审查了建立的范围检查。
回复率
---------------------------
由于从所有样本村庄和家庭中检索到了完成的问卷，因此无需调整非响应的权重，实现了100%的回复率。
抽样误差估计
---------------------------
计算了调查数据中抽样误差的量级。由于调查中的大多数估计都是加权比率的形式，因此比率的方差以比率估计的形式呈现。CCLS 2001报告的附录A详细讨论了抽样误差估计。