分层抽样框效率

相对于抽样误差理论而言 ,非抽样误差理论研究存在明显的差距 ,且至今尚未形成严密的体系。文章将抽样误差与非抽样误差结合考察 ,集中分析了分层抽样的误差问题。首先对忽略非抽样误差单纯研究抽样误差的研究方法提出了质疑 ,然后定义了“抽样框效率”这一中心概念 ,最后从理论上寻求一种利用样本信息来揭示抽样框效率状态的原则 ,并将这些原则贯彻于实践 ,给出了分层抽样框效率的强、弱、无三个标准 ,分别用于支持保留、修改、放弃原抽样框的三种决策 ,从而形成了完整工作流程。

On the efficiency of stratified sampling frames

There is still obvious distance between the theory of Non sampling Error and the one of Sampling Error, and so far there is not a complete theoretic system coming into being. Therefore, this article lays emphasis on Stratified Sampling and makes a analysis of Non sampling Error and Sampling Error as a whole. Firstly, it throws doubts on the methodology which simply concentrates on the Sampling Error. Secondly, it raises the key concept of Efficiency of sampling frames; finally, it puts forward three principles to determine the states of sampling frames. They are high efficiency, low efficiency and zero efficiency, representing the decisions to keep, to modify and to abandon the current sampling frames.