Milan Terek, The School of Management, Slovak Republic

Type of article: scientific article

Strany: 3 – 21

Abstract

In a statistical survey, it is necessary to determine such sample size, which ensures high significance of the obtained results in practice with sufficient reliability. It is also important that the sample size is not larger than necessary. The aim of the paper is to present the possibilities of determining the necessary sample size when estimating the proportion, mean and total using a confidence interval in statistical surveys, based on a random sample from both an infinite and a finite population, and to point out frequent errors occurring in this context. This can facilitate the selection of an appropriate sample size procedure and the correct interpretation of the estimation results.