对SPSS统计软件中Mcnemar检验结果的讨论

A discussion on the results of Mcnemar's test by SPSS statistics

目的比较手工计算与SPSS软件计算Mcnemar检验的结果,分析结果不一样的原因。方法选取相关书籍上配对设计资料的应用实例,用两种方法计算:一是利用Mcnemar检验公式进行手工计算,二是用SPSS统计软件计算,比较两种计算方法的结果有何不同。结果手工计算的结果为P=0.001,SPSS软件计算的结果为P=0.002,两种方法的结果有出入。结论 SPSS统计软件对配对设计四格表资料进行计算时采用的是一种精确检验,运用的是二项分布原理,虽然与手工计算的结果有些出入,但无论用哪种方法都不会影响到最终的统计推断。

Objective To compare the results of Mcnemar＇s test by hand calculation and SPSS statistics, and to explain reasons for the difference.MethodsApplication examples of paired design data in related books were selected, and two calculation methods were used： the first was a hand calculation using Mcnemar＇s test formula, and the second was a calculation by SPSS. Difference of the results done by the two methods were compared.ResultsThe result by hand calculation wasP=0.001, while the result by SPSS wasP=0.002. The results by the two methods were different.Conclusion SPSS statistics adopts an exact test in the calculation of 2 × 2 contingency tables with paired design data, in which a principle of bionominal distribution is applied. Despite the fact that the result differs from that of hand calculation, neither method will affect the final statistical inference.