论文部分内容阅读
通过对确切概率法、χ2方法、连续性校正法三种方法的计算机程序计算结果的比较,初步表明大样本,且理论频数大于5或7.5的某些四格表的显著性检验,采用χ2方法或连续性校正法会增加显著性检验的第一类错误或第二类错误的机率。因此大样本或理论频数较大的某些四格表,仍需用四格表的确切概率法(Fisher法)计算其概率。
By comparing the computer program results of the three methods of the exact probability method, the χ2 method, and the continuity correction method, it is preliminarily shown that a large sample, and a certain number of square tables with a theoretical frequency greater than 5 or 7.5, is used for the significance test. The χ2 method or the continuity correction method increases the probability of the first type of error or the second type of error in the significance test. Therefore, some four-square tables with large samples or larger theoretical frequencies still need to use the exact probability method (Fisher method) of the four-cell table to calculate their probabilities.