Sampling procedure
<p style="border:solid thin black;"> THE CLEANED AND HARMONIZED VERSION OF THE SURVEY DATA PRODUCED AND PUBLISHED BY THE ECONOMIC RESEARCH FORUM REPRESENTS 100% OF THE ORIGINAL SURVEY DATA COLLECTED BY THE PALESTINIAN CENTRAL BUREAU OF STATISTICS</p>
The methodology was designed according to the context of the survey, international standards, data processing requirements and comparability of outputs with other related surveys.
---> Target Population:
All Palestinians aged 10 years and over living in the Palestinian Territory, excluding persons living in institutions such as prisons or shelters.
---> Sampling Frame:
The sampling frame consisted of a master sample of enumeration areas (EAs) selected from the population housing and establishment census 2007, the master sample consists of area units of relatively equal size (number of households), these units have been used as primary sampling units (PSUs).
---> Sample Design
The sample is a two-stage stratified cluster random sample.
---> Stratification:
Four levels of stratification were made:
1. Stratification by Governorates.
2. Stratification by type of locality which comprises: (a) Urban (b) Rural (c) Refugee Camps
3. Stratification by classifying localities, excluding governorate centers, into three strata based on the ownership of households of durable goods within these localities.
4. Stratification by size of locality (number of households).
---> Sample Size:
The sample size was about 7,770 households in the 56th round and 7,818 households in the 57th round, and 7,819 households in the 58th round and 7,772 households in the 59th round. The total number of the households was about 31,179 households. The number of completed questionnaires was about 27,514 questionnaires and this was considered appropriate to provide estimations on main labour force characteristics in the Palestinian Territory.
The sample size in 1st quarter, 2010 consisted of 7,770 households, which included 29,999 persons aged 10 years and over (including 24,395 aged 15 years and over). In the 2nd quarter, the sample consisted of 7,818 households, which included 29,483 persons aged 10 years and over (including 24,118 aged 15 years and over). In the 3rd quarter, the sample consisted of 7,819 households, which included 28,479 persons aged 10 years and over (including 23,260 aged 15 years and over). In the 4th quarter the sample consisted of 7,772 households; which included 28,478 persons aged 10 years and over (including 23,288 aged 15 years and over).
---> Sample Rotation:
Each round of the Labor Force Survey covers all the 481 master sample areas. Basically, the areas remain fixed over time, but households in 50% of the EAs are replaced each round. The same household remains in the sample over 2 consecutive rounds, rests for the next two rounds and represented again in the sample for another and last two consecutive rounds before it is dropped from the sample. A 50 % overlap is then achieved between both consecutive rounds and between consecutive years (making the sample efficient for monitoring purposes). In earlier applications of the LFS (rounds 1 to 11); the rotation pattern used was different; requiring a household to remain in the sample for six consecutive rounds, then dropped. The objective of such a pattern was to increase the overlap between consecutive rounds. The new rotation pattern was introduced to reduce the burden on the households resulting from visiting the same household for six consecutive times.
Response rates
Errors due to non-response because households were away from home or refused to participate. The overall non response rate amounted to almost 11.8% which is relatively low; a much higher rates is rather common in an international perspective. The refusal rate was only 1.1% and the coverage rate was only 4.7%. It is difficult; however, to assess the amount of bias resulting from non response. PCBS has not yet undertaken any non-response study. Such a study may indicate, that non-response is more frequent in some population groups than in others. This is rather normal and such information is necessary to be able to compensate for bias resulting from non-response errors.