A Study Of Sample Rotation In Successive Sampling Survey

Since Jessen proposed the sample rotation of the successive sampling survey in the middle of the 20 th century, the research of the national statistical scholars of more than seventy years makes the theory of sample rotation developing rapidly. These scholars carried out a series of concrete research mainly for lots of practical problems existing in the process of successive sampling survey, mainly divided into two aspects: on the one hand is the pattern design of sample rotation, the other hand is the study of sampling estimation method. But the research mainly confined to the sample rotation of simple random sampling survey both at home and abroad, the study of the sample rotation problem of the successive sampling survey under complex sampling method is less.In view of this, this paper first summarizes the basic theory of sample rotation that is consistently used and scattered at home and abroad, which lays a theoretical foundation for the use of the following chapters. Secondly, this paper mainly discusses the sample rotation of stratified simple random sampling, the sample rotation of cluster sampling and the sample rotation of two stage sampling, on the basic of the sample rotation model under the way of simple random sampling when some auxiliary information can be available.This paper deduces their own sample rotation model with the auxiliary information available, given the regression composite estimator, the variance of the regression composite estimator, the optimal sample rotation rate, the optimal weight coefficient and so on. This enrich and perfect the sample rotation theory of successive sampling survey. And In the case of the effective information of auxiliary variables and the main variable indicators are highly correlated, the auxiliary variables can be effectively improved estimation methods, reducing the estimated sampling error and improve the estimation accuracy of the sampling.Therefore, the estimation precision of sample rotation under complex sampling can be improved more.