非参数检验方法概述

A review of nonparametric test

有些资料不是正态分布,或者分布情况未知,不能用有效的参数进行描述,此时需采用非参数法对总体的分布或分布位置进行检验。其优点是不考虑总体分布,适应性强,稳定性好,在临床研究中使用较多;缺点是损失了部分信息,检验效率较低,对于适合参数检验的资料,建议首选参数检验。本文对非参数检验进行讲解,以帮助临床医生掌握基本概念和方法。

For some data which have skewness distribution or unknown distribution, it is impossible to apply parametric test. Nonparametric test will be a good choice for this kind of data in testing the distribution of the population. The nonparametric test is widely used in clinical research by its goodness of stability and adaptiveness. However, the effectiveness is cutting down by losing some essential information. So, parametric test will be recommended when the data meet the requirements. This paper introduces the basic concepts and methods of nonparametric test for the physicians to understand this important tool.