中文摘要：

目的比较单向有序表的三种统计检验方法的功效。方法采用Monte Carlo模拟。结果模拟结果显示CMH法和K—W秩和检验对分布位置差异的检验有较高的功效，而对分布离散程度的变异却无能为力。结论分析单向有序表时应同时使用Pearson x^2和K—W秩和检验（或CMH法），并报告两种检验的统计结果。

英文摘要：

Objective To compare the power of Pearson X^2 test, CMH test and K-W rank sum test for one-way ordinal contingency table. Methods Monte-Carlo simulation was employed. Results Different from Pearson X^2 test, CMH test and K-W rank sum test are of high power for the background distributions with location-shlfted and breakdown when heteroscedasticity presented. Conclusion To use Pearson X^2 test and K-W rank sum test at same time for one-way ordinal contingency table is the best choice.