Calculation of the effect size when using nonparametric tests for comparing samples in psychological research using the R software

Abstract

Calculation of the effect size when using nonparametric tests for comparing samples in psychological research using the R software

Kupriyanov I.V.

Incoming article date: 24.05.2016

The limitations of statistical significance in psychological research become more and more obvious. On the contrary the effect size reporting in scientific papers is now a common practice as it shows the practical significance of the research. Various methods of computing the effect size exist, but many of them are not represented in popular statistical software. The using of the R statistical software for educational and scientific purposes is suggested, as it is free and open-source, and user functions for different purposes can easily be written. Four R functions for computing the effect size for rank comparison tests of samples are proposed.

Keywords: effect size, nonparametric statistics, rank tests, Mann-Whitney test, Wilcoxon signed-rank test, Kruskal-Wallis test, Friedman test, math statistics, R statistical software