Annual Employment Survey - 2002 - Sri Lanka

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

The Annual Employment Survey of Sri Lanka conducted by the Department of Labour has been designed to measure the levels and trends of the private sector and semi-government sector employment of the country over a period of time.
The main objective of the survey is to collect employment related information in the organised sector establishments of the country

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

National coverage

Although this is a national level survey, coverage of Nothern and Estern provinces was not at a sufficient level of response due to the prevailing situation in this area.

The coverage was also limited because of the fact that there was no complete list of establishments for the whole country at the time of the survey.

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

The survey covers establishments in both Public and Private sectors

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

The survey covered establishments with five or more paid employees in private sector and also semi-government institutions in public sector.

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

Sample survey data [ssd]

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

A complete list of establishments ( survey frame ) was available at the time of the first survey started in 1971. This list was compiled along with the list of buildings in the country prepared for the Census of Population and Housing 1971. There were about 22,500 establishments with 5 or more paid employees and thereafter this survey frame was updated annually, taking new registrations from the Employees Provident Fund lists.

Construction of indices:
Index numbers of employment are calculated with 1971 as the base year, and using the link relative method. Indices are computed for each major industry division,

Weighting of sample results:
The employment estimates are not weighted

Adjustments:
Non response : there are no adjustments for non response
Other bias : no adjustments have been made for any other bias.
Use of bench mark data : none
Seasonal variations : not relevant

Indicators of reliability of the estimates :
Coverage of the sampling frame : coverage is limited, as there is no complete list of establishments for the whole country.
Sampling error/Sampling variance : not relevant
Non response rate :

Non sampling errors : the main known source of bias is the limited coverage of the survey due to the lack of a complete list of establishments for the whole country

Sampling deviation
---------------------------

The coverage of the survey was badly affected due to the limiting factors such as the non availability of a complete list of establishments for the whole country and the nature of the method of data collection (postal survey method).

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

Mail Questionnaire [mail]

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

The questionnaire used in the Annual Employemnet Survey consists of three parts.

They are :
1. Establishment particulars
2. Classification of employees
3. Occupational category

The questionnaire was printed in Sinhala, Tamil and English.

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

An edit program was run to edit the raw data files.

Depending on the absence or presence of errors, the records in the raw data file were seperated into an error free file and a rejects file respectively. From the rejects file, a error report was generated indicating the reason for rejection of each record. The errors were checked and manually corrected on the report by referring to the questionnaire. Based on the corrected report, error records were updated by data entry operators.

The edit program was run again to edit the rejects file and the error free records were added to the error free file. If the rejects file was not empty the edit process was repeated.

摘要
---------------------------
由斯里兰卡劳工部执行的年度就业调查旨在衡量该国私营部门和半政府部门就业水平及趋势的变化。调查的主要目的是收集该国组织部门机构的就业相关信息。
地理覆盖范围
---------------------------
全国覆盖
尽管这是一项全国性调查，但由于该地区现存的局势，北部和东部省的回应程度并不充分。由于在调查时，全国范围内没有完整的机构列表，因此覆盖范围也受到了限制。
分析单元
---------------------------
调查涵盖了公共和私营部门的机构。
总体
---------------------------
调查涵盖了私营部门拥有五个或以上付费员工的机构，以及公共部门中的半政府机构。
数据类型
---------------------------
样本调查数据 [ssd]
抽样程序
---------------------------
在1971年首次开始进行的第一次调查时，已有完整的机构列表（调查框架）。该列表与为1971年人口和住房普查准备的全国建筑物列表一起编制。大约有22,500个拥有五个或以上付费员工的机构，此后该调查框架每年更新一次，从职工公积金名单中获取新的注册信息。
指数构建：
就业指数以1971年为基年，采用链式相对法计算。为每个主要行业部门计算指数。
样本结果加权：
就业估计值未加权。
调整：
非回应：未对非回应进行调整。
其他偏差：未对任何其他偏差进行调整。
使用基准数据：无。
季节性变化：不相关。
估计值的可靠性指标：
抽样框架覆盖范围：由于没有全国范围内机构的完整列表，覆盖范围有限。
抽样误差/抽样方差：不相关。
非回应率：
非抽样误差：主要偏差来源是调查的覆盖范围有限，这是由于缺乏全国范围内机构的完整列表。
抽样偏差
---------------------------
由于如全国范围内机构完整列表不可用以及数据收集方法（邮寄调查方法）的性质等限制因素，调查的覆盖范围受到了严重影响。
数据收集方式
---------------------------
邮寄问卷 [mail]
研究工具
---------------------------
年度就业调查中使用的问卷分为三个部分。
它们是：
1. 机构详细信息
2. 员工分类
3. 职业类别
问卷以僧伽罗语、泰米尔语和英语印刷。
数据清洗操作
---------------------------
运行了一个编辑程序以编辑原始数据文件。
根据错误的有无，原始数据文件中的记录被分为无错误文件和拒绝文件分别。从拒绝文件中，生成一个错误报告，指出拒绝每条记录的原因。通过参考问卷检查错误，并在报告上手动进行更正。根据更正后的报告，数据录入员更新错误记录。
再次运行编辑程序以编辑拒绝文件，并将无错误记录添加到无错误文件中。如果拒绝文件不为空，则重复编辑过程。