The article touches on the general issues, related to the use of artificially generated data in the process of development and implementation of mathematical-statistical methods in the sociological sphere. The main directions of application of simulations in sociology are briefl y described. The area of responsibility of sociologists in relation to mathematicians in the verification of mathematical-statistical methods and their integration into sociological scientifi c fi elds is outlined. The author’s classifi cation of mathematical and statistical methods is given depending on the need to use artifi cial data for their verifi cation. The classifi cation of simulations in the verifi cation of mathematical-statistical methods according to the degree of signifi cance of scientifi c projects is explained, while each class was accompanied by a real case within the framework of sociological problems.The experiment with frequency confi dence intervals by A. Kryshtanovsky, using artifi cial samples of sociological survey data, was described as a project of low signifi cance. The example of a scientifi c case of moderate signifi cance was the study by J. C. F. de Winter and D. Dodou of the prospects for using Student's t-test to analyze samples of observations, expressed in the ordinal Likert scale. The example of high-level scientifi c simulation was the author's own experience, associated with the introduction of Bayesian methodology into empirical sociology in the context of developing methods for analyzing eff ect sizes when conducting comparative binomial social experiments with binary data. The study raises questions on the requirements for publishing methodological studies using the generated data. The standards for publishing simulation studies are outlined, ranging from classical general conventions to the important standards, adopted in sensitive areas such as medical research.