Sampling Procedure
The sampling frame for the 2024 ILFS was derived from the inventory of Enumeration Areas (EAs) compiled during the 2022 Population and Housing Census (PHC). The sample was designed to provide representative estimates with a 95 percent confidence interval across three domains in Tanzania Mainland (TZM): rural areas, other urban areas, and Dar es Salaam; and 7 domains for Zanzibar, namely: rural, urban, Kaskazini Pemba, Kaskazini Unguja, Kusini Pemba, Kusini Unguja and, Mjini Magharibi.
The sampling design followed a stratified two-stage approach. In the first stage, EAs were selected within each stratum using Probability Proportional to Size (PPS). In the second stage, households were systematically selected from each of the previously chosen EAs. A total of 441 EAs were selected at the first stage in Tanzania Mainland and 152 in Zanzibar. At the second stage, 24 households were selected from each sampled EA, resulting in 10,584 households in Tanzania Mainland, 3,648 in Zanzibar, and a total of 14,232 households for the entire United Republic of Tanzania (URT).
Weighting
Fundamentally, the basic weight for each sampled household was calculated as the inverse of its overall probability of selection. This probability was determined through a multi-stage sampling process, where the chance of selecting a household was the product of the probabilities at each stage starting from the selection of the Enumeration Areas (EAs), down to the selection of households within each EA. By taking the inverse of this cumulative probability, the resulting weight ensured that each household's data contribution was appropriately scaled relative to its representation in the population. To accurately manage and compute these probabilities, detailed records were maintained in a structured Excel spreadsheet. This spreadsheet contained essential information from the sampling frame for each selected EA, including the number of households, sampling intervals, and selection probabilities. By tracking these elements at each stage of selection, the overall selection probability and corresponding sampling weight for each household could be systematically derived. These weights were then applied during data analysis to produce nationally representative estimates of labour force characteristics and trends.