抽样调查中最佳测量次数的确定

现场抽样调查中，由于测量误差的存在，使得所测变量实测值的方差增大，通过增加每个体的测量次数可以控制测量误差，但这样每个体调查费用增大。本文对测量信度Ｒ，每个体测量次数ｍ与相应所需的样本含量ｎｍ、调查费用Ｔｎ的关系进行了探讨，并介绍了如何根据Ｒ，及每个体测量费用占其总费用构成比Ｃ，确定最佳测量次数ｍ值，以达到最佳控制调查费用的目的，这对我们在大型现场调查中进行经济效益分析具有重大的理论指导意义。

The determination of the Optimal Number of Replicate Measurements per Subject in Sampling Investigation

Investigators are always annoyed with the measuring error because it makes the variance of the observed value (X) on some variables larger. To acchieve the necessary measuring results, we should increase the sample size, or increase the number of measurements per sbject. In each case the study cost is certainly increased. Which method is more economic? After detailed discussing the relationship of the reliability of measurements (R), the number of measurements per subject (M) and the corresponding necessary sample size (NM) and the total cost of the study, we suggest a method to determine the optimal number of replicate measurement per subject in sampling investigation according to the reliability and the ratio of the cost of obtaining a measurement on a subject to the sum of all costs involved in having a subject under study, which is certainly significant for us to do economic benefit analysis in large sampling Investigation.