参数与非参数统计分析的过渡—秩变换

Rank transformations--the connection between nonparametric and parametric statistics

The purpose of this paper is to present the relationship between nonparametric and parametric analysis by means of rank transformation. It is shown that in case of large sample, the resulting statistics getting from Wilcoxon rank test, Kruskal-Wallis and Friedman rank test are equivalent to the ratio of the sum of squares for treatment divided by mean square for the total variability calculated by ranks in the manner of the analysis of variance. It is also suggested that this method can be extended to factorial design experiments, and a detail procedure is given.