Sampling Procedure
The Labour Force Survey (LFS) employs a stratified two-stage sampling design based on the 2001 Population Census. The country was divided into 80,787 enumeration areas (EAs), from which a Master Sample of primary sampling units (PSUs) was constructed. Small EAs with fewer than 25 households were excluded, while those with 25–99 households were combined to form PSUs, and large EAs were split into multiple PSUs. This design ensures representativeness at provincial, metro/non-metro, and geography-type levels (urban formal, urban informal, farms, and tribal). The overall sample size for the Quarterly Labour Force Survey (QLFS) is approximately 30,000 dwellings, drawn systematically from PSUs selected with probability proportional to size.
The current sample comprises 3,080 PSUs, divided into four rotation groups to facilitate quarterly updates while maintaining representativeness. Each rotation group mirrors the distribution of the full sample, allowing systematic replacement across survey waves. In the first stage, PSUs are selected using probability proportional to size, and in the second stage, dwelling units are sampled systematically within each PSU. This rotation and stratification framework ensures that the LFS produces reliable, nationally representative labour market statistics while balancing efficiency and continuity in data collection.
Researchers should refer to the VAS 2024 Metadata document for more information on sampling.
Weighting
The sampling weights for data collected from the sampled households are constructed so that the responses can be properly expanded to represent the entire civilian population of South Africa. The weights were constructed to account for several factors, including original selection probabilities, adjustment for non-response, and benchmarking to known population estimates from the Demography Division of Stats SA.
Researchers should refer to the VAS 2024 Metadata document for more information on weights.
