配对四格表资料χ2检验在 SPSS 和 SAS 软件中的实现

The implementation of chi-square test of paired design in SPSS and SAS statistical software

目的：通过实例来介绍配对四格表资料的卡方检验在SPSS和SAS统计分析软件中如何操作，输出结果的区别及正确的解读，为临床及相关科研人员提供可以借鉴的方法。方法使用SPSS和SAS统计软件，对配对四格表资料的卡方检验进行操作和分析。结果本文以两种方法诊断肺癌的检测结果为例题，根据基本公式算得统计量χ2＝4．92，P＜0．05，得出两种诊断方法的诊断结果存在差异。在SPSS软件的分析结果中只有P值，SAS软件的分析结果可以给出统计量和P的确切值，但统计量χ2＝6．23，与基本公式计算的结果不同，经过自编程序运行后，我们得到了与基本公式相同的结果。结论 SPSS和SAS统计软件是医学统计学数据分析的常用教学软件，本文通过实例详解，得出在结果的输出方面两种软件有很大区别，但这并不影响我们得出相同的统计结论。 SPSS软件操作简便，SAS软件编程比较麻烦，以及SAS软件在处理此类资料上存在默认不校正的问题，因此，在实际应用中，我们可以根据情况自行选择。

Objective To introduce the operation of chi-square test of paired design in SPSS and SAS statistical software by an example, to differentiate and correctly interprete outcomes, and then to provide a reference for clinical investigators.Meth-ods With SPSS and SAS statistical software, the chi-square test of paired design was analyzed.Results In this article, we se-lected the results of two methods for the diagnosis of lung cancer, and calculated by corrected formula：χ2=4.92, P〈0.05, and therefore, we could conclude that the difference of two diagnostic methods had statistical significance.In the outcome of SPSS, we could only get P value, and we could get exact value of statistic and P from the result of SAS, but the result was different from that of corrected formula.After we ran self-compiled program, we got the same result.Conclusion SPSS and SAS are the com-mon software which are used to analyze the data in medical field.We find that it is different in the output of results, but it does not affect to draw the same conclusion.SPSS is simple and SAS is complex.We can have our choice in practical applications.