摘要
目的 比较t检验与秩和检验检验完全随机设计两组资料的功效。方法 用SAS软件编制电脑实验程序 ,模拟研究和比较不同总体条件下两组秩和检验与t检验的检验功效。结果 若总体分布对称 ,小样本时t检验功效较高 ,大样本时两种方法功效相似 ;总体非对称分布时秩和检验的功效高于t检验。结论 当样本量足够大时 ,可以用秩和检验代替t检验。
Objective To compare power of the t-test and Wil coxon Rank-Sum test.Methods An investigation evaluates the comparative power of the independent samples t-test and its nonparametric counterpart, the Wilcoxon Rank-Sum test using SAS program.Results The t-test was more powerful under relatively symmetric distributions only with small sample sizes. When sample size is large enough, the power of the independent samples t-test and Wilcoxon Rank-Sum testis consistent. Under distributions with extreme skews, the Wilcoxon Rank-Sum test held large power advantages. Especially the distributions consist of heavier tails or extreme skews ,with large sample sizes.Conclusion When sample size is large enough, the Wilcoxon Rank-Sum test is recommended.
出处
《中国卫生统计》
CSCD
北大核心
2004年第1期10-13,共4页
Chinese Journal of Health Statistics
