A sampling project applied in a survey for estimating the vaccination coverage in Araraquara County, S. Paulo State (Brazil), in 1983, is described. A target population of children aged 12-13 months was divided into the strata: children living in private dwellings (i.e. single-family houses) and those living in collective dwelling places (orphanage, asylum). R.H. Henderson and T. Sundaresan's method, proposed by the World Health Organization (WHO) for use by the Expanded Programme on Immunization (EPI) was applied to the first stratum. This method consists of the random selection of 30 partial areas with a degree of probability proportional to the respective partial populations in each area chosen; a starting point ("household") is selected at random within each selected area; the selection of 7 children within each area begins with the first household and continues according to fixed rules. The sample of the first stratum was taken from areas in both urban and rural zones. In the urban zone, after the selection of the area and before the selection of the starting point, an intermediate stage, which consist in the selection of a block with probability proportional to its estimated population, was introduced. A sample was selected form the second stratum with the same sampling fraction used in the first stratum. The ratio-estimate is proposed as the estimator of the vaccination coverage rate and the formula for the calculation of the standard error is presented.