中文摘要：

目的 为实际需要的样本轮换下两阶段抽样连续调查提供科学的调查方法与统计公式；为制订某核电站职工辐射防护措施提供依据。方法 采用数理统计学的理论方法推导统计量的计算公式；采用本文研究的样本轮换下两阶段抽样连续调查的统计方法，对某核电站职工的白细胞数进行了连续三年的调查分析；采用SAS编程模拟调查分析100个样本，对本文研究的调查方法及其统计公式作信度与效度评价。结果 对样本轮换下两阶段抽样连续调查，推导出总体均值的估计量及其方差与估计方差的计算公式；该核电站职工两阶段抽样均轮换样本的2010年、2011年白细胞数总体均值的估计量分别为5.88、5.84（10^3/mm^3），其标准误分别为0.247、0.255，与一般成人白细胞均数的差异具有统计学意义；100个总体均数的95%可信区间均包含模拟总体均数。结论 本文研究的样本轮换下两阶段抽样连续调查的统计方法具有较好的理论与实际意义以及较高的效度与信度；该核电站职工的白细胞数偏低，应引起相关部门的高度重视。

英文摘要：

Objective To provide a scientific survey method and formulae of sample rotation under two-stage sampling, and to offer scientific basis for the relevant departments to set the strategy on radiation protection. Methods We derived the statistic formulae with mathematical statistics method. Sample rotation under two-stage sampling was applied to successive survey on the number of white blood cells （WBC） three years. SAS programming to simulate 100 samples was taken to make validity and reliability assessment on the survey method and its statistical formulae. Results We derived the formulae of estimation of population mean, variance and estimated variance for the sample rotation under two-stage sampling. After two-stage sample rotation, population means of nuclear workers’ WBC were respectively 5.88 and 5.84 （10^3/mm^3） in 2010 and 2011, and corresponding standard errors were respectively 0.247 and 0.255, which were of statistical differences compared with healthy adults. Simulated population means was totally included in 100 population mean 95% confidence intervals. Conclusion Sample rotation under two-stage sampling and its formulae have achieved well theoretical and practical significance, which are also of high validity and reliability. For the nuclear workers, numbers of their WBC were lower than healthy adults in 2010 and 2011, which should be paid more attention to.