使用基于约束性极大似然估计的统计量,对不完全2×2列联表中的风险比进行置信区间构造.导出在一定置信水平下控制置信区间宽度的样本量公式.模拟研究和实例分析验证了所给方法的合理性.
Confidence interval construction of risk ratio in an incomplete correlated 2×2 table is considered with the constrained-maximum-likelihood-estimation-based statistics.Sample size formulae that control the width of confidence interval with a certain confidence level are proposed.Simulation results and an numerical example show the validity of the methods.