The paper examines the problem of using mathematical statistics methods in pedagogical research.
It identifies and describes the possibilities of using different methods of analysis of nominal data
in processing the results of pedagogical experiments. The authors analyzed the frequency of use of
statistical criteria in 100 publications in journals included in Web of Science and the Russian Index
of Scientific Citation. The criteria for selecting publications were (1) year of issue; (2) reviews of
manuscripts; (3) free access to the publication from the publisher; (4) availability of quantitative
data on test results before and after pedagogical impact; (5) conclusions about the effectiveness
of pedagogical impact. It was found that 45% of publications did not use mathematical statistics
methods to test scientific hypotheses. Of the mathematical statistics methods used by the authors,
the T-Test and Chi-square Test are more often used. For the study example of developing conflict
readiness in future teachers, it presents the possibilities of using: Chi-square Test, Fisher’s Exact Test
(Phi test), Cramer 's V Test and Somers’ d Test.