用传统显著性检验方法进行等效性检验的规律研究

目的:研究传统显著性检验同等效性检验的联系规律,并探讨在传统显著性检验结果为P>α时,依据等效性界值结合本研究结果下等效结论的可能性。方法:在MATLAB软件上编程进行模拟研究。其中,编写程序的正确性均通过了SPSS13.0软件和EquivTestTM2.0软件的验证。结果:得到了各种情形下基于传统显著性检验P值进行等效性判定的标准,以及不同判定标准下作等效性判定的判错率,并以方便查询的表格形式表达出来。判错率与P值间呈三次模型关系。结论:各医药专业研究人员不必专门系统地学习等效性检验的理论,而可据本研究结果直接利用传统显著性检验方法作出是否等效的判断。

Laws of Conducting Equivalence Test Using Traditional Significance Test Method

OBJECTIVE: To explore the relation laws between traditional significance test and equivalence test,and to probe into the probability to make an equivalence conclusion when P value is larger than α in traditional significance test based on the equivalence boundary and the study results.METHODS: By means of simulating with programs,the research was conducted on MATLAB software packages.All the programs concerned in tests were verified by SPSS 13.0 and EquivTestTM 2.0.RESULTS: The standards of making equivalence conclusions based on the P values of traditional significance test as well as the rates of erroneous judgment under different Pvar values were obtained.All the results were listed in forms for briefly inquiring.The cubic model could indicate relations between P values and rates of erroneous judgment.CONCLUSION: It is unnecessary for clinical researchers to pay special efforts to learn theories about the equivalence test;instead,they can depend on the traditional significance test and results of this study to decide whether it is equivalent or not.