Botswana - Botswana Labour Force Survey 2005-2006

SAMPLING FRAME
In general the 2001 Population and Housing Census, undertaken in August, is the Sampling Frame on which sample selection for the Survey Programmes are based. The census result gives information on population, number of household at Locality, Enumeration Area (EA), village and district/town levels. Also given for each EA is information on ecological zones in rural areas.

The Sampling frame was defined and constituted by all Enumeration Areas (EAs) found in three geographical regions viz. (i) Cities & Towns (ii) Urban Villages, and (iii) Rural Districts as defined by the 2001 Population and Housing Census.

Being a two-stage design, two frames were required one for each stage.

The sampling frame for the first stage based on the 2001 Population and Housing Census. This comprised the list of all Enumeration Area (EA) together with number of households. In the census the EAs were frames of manageable size (in terms of dwellings/households).

The sampling frame for the second stage was produced only in the selected EAs. Before the beginning of the survey interviews, the field teams listed all private habitable dwellings/households in their EAs. Thus the number of occupied households in the selected EA served as sampling frame for that EA.

The frame for the Botswana Labour Force Survey 2005/6 consisted of 4,143 EAs being the total number of Enumeration Areas (EAs) delineated during the 2001 Population and Housing Census.

STRATIFICATION
When national level estimates are the main focus a type of stratification that is simple to implement and highly efficient is implicit stratification. It is a form of geographic stratification, which when used together with systematic pps sampling automatically distributes the sample proportionately into each of the nation's administrative subdivisions, as well as the urban and rural sectors.

Creation of strata is dictated by two principal criteria. These include a need to:
i. provide estimates for each major region of the country.
ii. increase precision

Thus, stratification variables included cities/towns and administrative districts. Apart from national and rural estimates, the Government, which is the main user of CSO data, requires accurate estimates for all regions for planning and monitoring of development projects. Stratification was therefore undertaken such that all districts and major urban centres become their own strata. With regard to increase precision consideration was also given to group EAs according to ecological zones in rural districts and according to income categories in cities/towns.

Geographical stratification along ecological zones and income categories was expected to improve the accuracy of survey data in view that homogeneity of the variables was relatively high (implicit stratification).

There are five major rural ecological zones, namely:
-Village,
-Lands
-Cattle Post
-Freehold Farms
-Mixture of Land and Cattle Post

During the delineation of the maps, each EA was associated with unique ecological zone and thus, grouping the EAs into respective zones was not a problem. To facilitate the selection according to the stratification variables and EAs were listed in some order, for example starting cattle post, then farms etc. in case of rural areas.

Note: See detail sampling procedure description in final report


CALCULATION OF SAMPLING WEIGHTS
Being a multistage design, it follows naturally that the sample selected at each stage represents (or is assumed to) the respective population. The fundamental assumption was that units selected at each stage were similar to those not selected, in respect of characteristics of interest. In the treatment of unit for the non-response the assumption that the responders were similar to non-responders though should not be always taken for granted.

The weights of the sample are equal to the inverse of the probability of selection. Therefore the sampling probabilities at first stage of selection of EAs including probabilities of selecting the households were used to calculate the weights.

Note: See detailed sampling weight calculations in the final report.