Sampling Procedure
The sample was drawn from a census conducted by the Kenya Central Bureau of Statistics (CBS) of nearly 2000 formal manufacturing firms employing more than 250000 full-time employees. In order to ensure representation of all types of firms, the sample was stratified across location, sub-sectors, and size in 148 clusters.
The survey covered seven sub-sectors in five urban areas: Nairobi, Mombasa, Eldoret, Kisumu and Nakuru.
Nine manufacturing sub-sectors were: Agroindustry, Chemicals/Paints, Construction Materials, Furniture, Metals, Paper/Publishing/Printing, Plastics, Textile/Leather, and Wood. Four size classes were used: Small (11-49 employees), Medium (50-99 employees), Large (100-499 employees), and Very Large (500 and more employees).
368 firms were selected randomly from the clusters, representing roughly 20 percent of all formal firms. Several firms, often non-African and foreign ones, refused to be interviewed. Whenever possible, these firms were replaced with "new" firms having the same characteristics as the ones that refused. Due to the high rate of
refusal, however, the replacement strategy was only partly successful and, in the end, 284 firms completed a survey.