On the statistical indicators of the effectiveness of teaching methodologies

Abstract

The article is devoted to the problem of carrying out a proper educational experiment. A critical analysis of the conventional approaches to an educational experiment is given. The author argues that the majority of methods, assessing the results of educational experiments, were borrowed from sociology and psychology which, in turn, led to the misinterpretation of the results of these experiments. The criticism of the author is primarily aimed at the incorrect use of central tendency measures and the selection of tests for checking the probability of significance of samples. The qualitative approach, based on the percentile values, was proposed as one of the most relevant results of an experiment. The lack of universal measure with could allow comparing results of multiple educational experiments and meta-analyses was argued. The term «effective learning hours» was coined. The methodology of defining «effective learning hours» and corresponding them to the levels of the acquisition was proposed.