The exact unconditional z-pooled test for equality of two binomial probabilities: optimal choice of the Berger and Boos confidence coefficient
The exact unconditional z-pooled test for equality of two binomial probabilities: optimal choice of the Berger and Boos confidence coefficient
复制标题
DOI：
10.1080/00949655.2011.579969
复制
发表时间：
2012-01-01
影响因子：
1.2
通讯作者：
Bakke, Oyvind
中科院分区：
文献类型：
--
作者：
Lydersen, Stian;Langaas, Mette;Bakke, Oyvind
用于比较两个二项式概率的精确无条件检验通常比费舍尔精确检验等条件检验更强大。它们的功效可以通过 Berger 和 Boos 置信区间方法进一步提高，其中通过将 H-0 下的常见二项式概率限制为 1 - gamma 置信区间来找到 p 值。我们研究了具有平衡和不平衡样本量以及显着性水平 0.05 和 0.01 的各种情况下精确无条件 z 池检验的平均检验功效。详细结果可在网上获取。在值 10(-3)、10(-4)、...、10(-10) 中，值 gamma = 10(-4) 在我们研究的所有情况下给出了最高功率或接近最高功率，并且可以作为最佳 gamma 的一般建议给出。
Exact unconditional tests for comparing two binomial probabilities are generally more powerful than conditional tests like Fisher's exact test. Their power can be further increased by the Berger and Boos confidence interval method, where a p-value is found by restricting the common binomial probability under H-0 to a 1 - gamma confidence interval. We studied the average test power for the exact unconditional z-pooled test for a wide range of cases with balanced and unbalanced sample sizes, and significance levels 0.05 and 0.01. The detailed results are available online on the web. Among the values 10(-3), 10(-4), ... , 10(-10), the value gamma = 10(-4) gave the highest power, or close to the highest power, in all the cases we looked at, and can be given as a general recommendation as an optimal gamma.