基于多重抽样框的校准估计方法研究

The Research on the Calibration Estimation Method Based on Multiple Sampling Frames

在现代抽样调查中,校准估计方法能够通过有效利用辅助信息来提高估计量的精度,多重抽样框抽样调查则不仅可以解决单一抽样框覆盖不全的问题,还可以节约抽样设计阶段的成本。本文将这两种现代抽样估计与设计方法进行结合,将校准估计方法引入到基于多重抽样框的抽样调查体系中,在节约调查成本的同时,还能够提高估计量的精度。本文首先按照分离抽样框与组合抽样框估计方法的分类思路,对传统多重抽样框估计方法进行系统梳理;然后在最短距离法校准估计的分析框架下,按照调查时所能掌握辅助信息的具体情况,给出了两类多重抽样框估计情形下的各种不同形式的校准估计量;随后数值分析的比较结果也表明在多重抽样框中校准估计量的估计效率明显优于传统估计量;最后对本文研究进行总结,对我国抽样实践中应用这套抽样估计方法体系进行了展望。

In modern sampling surveys, calibration estimation method can improve the accuracy of estimator by effectively using auxiliary information. The sampling surveys through multiple sampling frames can solve the problem of incomplete coverage in single sampling frame and save the cost of sampling design phase.This paper combines these modern sampling estimation and design method and introduces the calibration estimation method into the sampling surveys based on multiple sampling frames. At the same time,the investigation cost can be saved,and the estimator＇s accuracy can be improved. Firstly,it sorts the traditional multiple sampling frames estimation method according to the divided sampling frames and combined sampling frames estimation method. Then,under the analytical framework of the shortest distance method calibration estimation,it gives the two calibration estimators of multiple sampling frames estimation according to the situation of the auxiliary information grasped. The results of numerical comparison and analysis show that the calibration estimators are superior to the traditional estimators in multiple sampling frames. Finally, it summarizes all results and looks forward to the application of sampling practice in China.