In the statistical analysis it is sometimes necessary to determine sample size. It is not required if the results are not collected by means of statistical apparatus, and it serves only to present the observed phenomena or mileages. In the presented material some ways to determine sample size and sampling are given. Sampling from unknown and known set, drawing sample from strata, as well as multistage sample, systematic and in beams is characterized. Also comments on the costs of sampling and the proportions in the sample are presented.