Mathematical statistics has been used in various areas because of its accurate and objective results, and relatively small errors. Using statistics in research involves the collecting of samples, or a set of independent samples representing a whole group. There remain, however, cases where sample selection is not unbiased, the samples do not accurately represent the whole population, and then the results are undesirable and even contrary to the law of practice. In this paper, we present our study of two random sampling methods: simple random sampling and stratified random sampling. While stratified random sampling costs and is a complex and time-consuming process, its accuracy is higher than that of simple random sampling.