Household Income and Expenditure Survey - 1990-1991 - Sri Lanka

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

This survey provides information on household income and expenditure to be able to measure the levels and changes in the living condition of the people and to observe the consumption patterns .

Key objectives of the survey
- To identify the income patterns in Urban, Rural and Estate Sectors & provinces.
- To identify the income patterns by income levels.
- Average consumption of food items and non food items
- Expenditure patterns by sector and by different income levels.

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

National coverage.

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

Household, Individuals

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

For this survey a sample of buildings and the occupants therein was drawn from the whole island

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

Sample survey data [ssd]

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

Sample Design
A two stage stratified random sample design was used in
this survey. Bach domain for which separate estimates were
required was made a separate stratum. As such each sector
(Urban, Rural, Estate) within each district was considered as a
separate stratum.

Sample Size
In this design the first stage units (FSU) were the census
blocks prepared at the 1981 Census of Population and the
second stage units (SSU) were the housing units. It was decided
to select 10 housing units from each selected census block.
Thus, a first stage sample of about 2,500 census blocks have to
be selected from the entire island.

Sample Allocation and Method of Selection
The allocation of 2,500 census blocks to each district
was made proportional. to the square root of the population· ( as
at 1981 Census ) in that district. These values have been rounded
to multiples of twelve. ( Refer page 5 of the final report attached in the external resources section)
It was decided to over sample the urban sector in each district
in comparison to the rural and estate sectors with the objective
of allocating roughly by one-third of the total sample.
Allocation between the rural and estate sectors in each
district was proportional to 1981 population. Within each
stratum, the assigned number of FSU's were selected with probability
proportionate to size (using the Census or adjusted housing
unit counts ) with replacement.
The lists of census units prepared for the Census of Population
and Housing 1981, of each selected block were updated to include
new ·housing units and to exclude ones which are no longer
in existence. This updating operation was also staggered over a
period of twelve months starting from May 1990 to April 1991.
For each FSU, updating was done about one month prior to the
scheduled interviewing.

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

Non-Response
Of a sample of 25080 housing units selected only 19,401
households were covered in the survey. Northern and Eastern
provinces were excluded in the survey due to the prevailing conditions
in the area.

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

Face-to-face [f2f]

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

Questionnaires

The survey schedule was designed to collect data on household basis and separate schedules were used for each household (identified according to the definition of the household) within the housing units selected for the survey. The survey Schedule consists of three main sections .

1. Demographic section
2. Expenditure
3. Income

The demographic characteristics and usual activities of the inmates belonging to the household are reported in the Demographic section of the schedule and close relatives temporarily living away are also listed in the section. Expenditure section has two sub sections to report food and non-food consumption data separately. Expenditure incurred on their own decisions by boarders and servants are recorded in the sub section under the expenditure section. The income has seven sub sections categorized according to the main sources of income.

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

Out of 19401 households 95 percent the households were fully completed while
0.2 percent have refused.
Sector wise completion rates show that Urban sector had 93
percent, Rural sector 96 percent and the Estate sector 97 percent.

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

Refer the pages 7 to 10 of the final report attached in the External Resources section.

摘要
---------------------------
本调查旨在收集家庭收入与支出数据，以便衡量民众生活水平及其变化，并观察消费模式。

调查的主要目标
- 识别城市、农村和住宅区的收入模式及各省份的收入模式。
- 按收入水平识别收入模式。
- 食品及非食品类平均消费。
- 不同收入水平及各行业支出模式。

地理覆盖范围
---------------------------
全国范围。

分析单元
---------------------------
家庭，个人。

调查范围
---------------------------
本调查从整个岛屿抽取了部分建筑及其居住者作为样本。

数据类型
---------------------------
样本调查数据 [ssd]。

抽样程序
---------------------------
抽样设计
在本次调查中采用了两阶段分层随机抽样设计。对于需要单独估算的每个领域都构成了一个单独的层。因此，每个区域（城市、农村、住宅区）内的每个区域都被视为一个单独的层。

样本量
在本设计中，第一阶段单元（FSU）是1981年人口普查准备的普查区块，第二阶段单元（SSU）是住宅单元。决定从每个选定的普查区块中选出10个住宅单元。因此，需要从整个岛屿中选出约2,500个普查区块作为第一阶段样本。

样本分配和选择方法
决定在每个区域中对城市部门进行超抽样，以农村和住宅部门相比，目标是分配约三分之一的总样本量。农村和住宅部门在每个区域之间的分配与1981年人口成比例。在每个层内，按大小概率（使用普查或调整后的住宅单元计数）选择分配的FSU数量。

抽样偏差
---------------------------
非响应
在选出的25,080个住宅单元样本中，只有19,401个家庭参与了调查。由于该地区的现有条件，北部和东部省份被排除在调查之外。

数据收集方式
---------------------------
面对面 [f2f]。

研究工具
---------------------------
问卷。

调查安排旨在收集基于家庭的数据，并为调查选定的每个住宅单元内的每个家庭（根据家庭的定义进行识别）使用单独的安排。调查安排包括三个主要部分。
1. 人口统计部分
2. 支出
3. 收入

人口统计部分报告了属于家庭的居民的人口统计特征和日常活动，并在该部分中列出临时居住在外地的近亲。支出部分有两个子部分，分别报告食品和非食品消费数据。在支出部分的子部分中记录了由寄宿者和仆人自行决定产生的支出。收入部分分为七个子部分，按收入的主要来源进行分类。

响应率
---------------------------
在19,401个家庭中，95%的家庭完成了调查，0.2%的家庭拒绝了调查。

抽样误差估计
---------------------------
请参阅外部资源部分附上的最终报告的第7至10页。