Small Area Estimation Method In Dual Frame Survey Sampling Design And Application

With the rapid development of China’s economy and the intensification of market competition,the changes in the micro-subject units in the sample survey have become increasingly frequent.It is difficult to achieve a complete coverage of the target population by using a traditional single list frame for sampling surveys.Insufficient sampling frame coverage is a serious defect in sampling design.At present,the vast majority of developed countries adopt dual sampling frame design to solve the problem of insufficient coverage of single sampling frame.China has begun to explore the application of dual sampling frame surveys.Driven by the frequent changes in sample survey subjects and the modernization reform of the statistical survey system,dual sampling frame surveys have received higher attention,in the face of the China’s special government hierarchical management model,the Multi-level estimation of statistics at all levels of government needs to be considered.Small area estimation is an effective means to solve this problem.Based on this background,this study introduces the small area estimation method into the dual sampling frame survey system,that is,the small area estimation method is applied in the dual sampling frame survey to analyze the statistical survey data of governments at all levels.The research objective of this study is to combine the Chinese government’s multi-level management model for the dual sampling frame survey,and to deeply study the small area estimation method under the dual sampling frame survey,so as to carry out the dual sampling frame survey for China and provide the statistical survey required by the government at all levels method selection.This study focuses on the small area estimation method and application under the dual sampling frame.By systematically reviewing consequential research literature on dual sampling frame surveys and small area estimation methods,as well as the practical experience of developed countries,the technical details of dual sampling frame design and small area estimation methods are summarized.Under simple random sampling and two-stage sampling,the construction process of dual sampling frame small area estimator is systematically studied,and the statistical properties of the new method and its applicability in practice are discussed by using random simulation method and empirical research method.Based on the above ideas,the main research contents of this study are as follows:(1)According to the ideas of separate sampling frame and combined sampling frame estimation in dual sampling frame design,under the assumption of simple random sampling,according to the different situations of auxiliary information in dual sampling frame,different forms of small area estimator are constructed.On this basis,it is proved that the small area estimator is unbiased,and the Pseudo-population Bootstrap method is applied to estimate the variance of the estimator.In addition,in view of the problem of obtaining the summary value of auxiliary variables in small area in each separation frame,the idea of synthetic estimators is used to estimate them by using the information in large area.Finally,different forms of small area estimators are simulated by statistical simulation methods.The results show that the small area estimators of the dual sampling frame have better estimation results than the HT estimators,and the Pseudo-population Bootstrap method can effectively estimate the variance estimate of the estimators.(2)Aiming at the two-stage sampling commonly used in the practice,according to the information of the auxiliary variables in the small area in different stages,different forms of dual sampling frame small area estimators are constructed,and on this basis,the two-stage sample is discussed.The construction process of small area estimator under two-stage sample rotating sampling is discussed,in order to provide a theoretical basis for the small area estimation method based on two-stage sampling under dual sampling frame in sampling survey practice.The simulation results show that the small area estimator under two-stage sampling and two-stage sample rotation sampling can effectively estimate the target variable in the small area.(3)Taking the China’s "agriculture,countryside and farmers" sample survey as an example,this study conducts an empirical application research on the small area estimation method under the dual sampling frame.Combined with the actual situation of the "agriculture,countryside and farmers"sampling survey,the design scheme of the dual sampling frame combining the list frame and the regional frame is given.Using the micro sample data of farmers in the third national agricultural census,the small area estimation under the dual sampling frame is applied.The information of farmers in the sample counties was estimated.The empirical research results show that the relative root mean square error of the small area estimator is small,and the accuracy of the estimator is high,but the total value of the target variable is underestimated.At present,China has entered a stage of high-quality development,which puts forward higher requirements for government statistics at all levels.Facing the frequently changing survey subjects and the multi-level management model of the government,this study introduces the small area estimation method into the dual sampling frame survey,systematically studies the small area estimation method under the dual sampling frame,and demonstrates the statistical properties and practical applicability of the new method using statistical simulations and empirical studies.The research content of this study improves the investigation theory of dual sampling frame,and provides the support of statistical investigation theory and method for promoting the modernization and reform of the statistical investigation system in the face of frequently changing microscopic investigation subjects.