The Research On Theory And Its Application Of Split Panel And Rotating Panel

With the rapid development of experimental sciences and other disciplines, it is fundamental significant for researchers to access to information scientifically and reasonably. Sampling survey is a kind of scientific method of collecting information. The theory and practice of sampling survey have developed rapidly and been applied to many fields widely. Panel samples and repeated samples are two kinds of common sampling designs, and they have their own advantages. But the disadvantages gradually highlight in practice.On the one hand, both panel samples and repeated samples can’t cut cost by adjusting the sample structure. To every research field, the cost of sampling survey is limited, and exorbitant cost can lift the budget of sampling survey, which affects the effect of sampling survey. On the other hand, because of the longer sampling time interval, serious problems such as unrepresentative problem that are caused by sample attrition and aged sample widely exist in panel samples, which affect the authenticity and reliability of information.In the sampling design of rotating samples and split panel, part sample continually update with sampling periods changing, so rotating and split sampling survey can relieve sample attrition and aged sample. Split panel consists of repeated sample and panel sample or rotating panel and panel sample. The structure decides that split panel combines the advantages of rotating samples, panel samples and repeated samples together and has the advantages of application flexibility and low cost. For practical research, researchers choose the proportion of panel samples in split panel according to actual situation, and meet the needs of research by using limited budgets. At the same time, we can choose part sample of split panel to use again. So according to the current problems such as sample attrition, aged sample and increasing sampling budget, this paper breakthrough the traditional sampling theory, extend the research of panel samples and repeated samples to that of rotating panel samples and split panel samples, which can improve the quality of data, decrease the cost, and meet the needs of efficient and practical data. On this basis, the paper study deeply on four aspects of rotating panel samples and split panel samples:Firstly, to use finite information to obtain more accurate estimate during sampling survey, based on the analysis of variance model, this paper discusses the optimal choice among rotating panel samples, panel samples, split panel samples from three different aspects. First, we get the variance-covariance matrix of the best linear unbiased estimation by mathematical deduction, and then we analyze the optimal choice on every form of parameters by numerical simulation.Secondly, from a new perspective, based on the analysis of variance model this paper discusses the optimal design of the first form split panel when ignoring the cost. And we study the optimal choice and the decision of sample size on three usual form estimations of the analysis of variance model during practical process. And then we transfer the optimal design of the first form split panel when considering the cost to the nonlinear inter program. Finally, the simulated annealing algorithm is designed to solve the nonlinear integer program and OATD and FD methods are used to analyze and compare the influence of the different selection of parameters to the implementation of the algorithm. Analysis demonstrates that the efficiency and precision of these methods can be improved clearly with the proposed suggestions.Thirdly, this paper systemizes the estimation method on the rotating panel samples in foreign countries. On this basis, we analyze the feasibility of FGLS method on one half rotating panel by Monte Carlo simulation. It clearly shows that FGLS method induced overconfidence when used for one half rotating panel data and the true sampling variability is distorted seriously. According to the deficiency, we propose the improvement ideas and present the robust estimation. Under the conventional assumption of variance component and stationary conditions, we improve the asymptotic properties of the robust estimation. Simulation studies show that compare to the FGLS estimation, from three degrees our new method can decrease the overconfidence on the premise of remaining A-Bias and MSE the same basically.Fourthly, we analyze comprehensively the problems on rotating panel survey of new rural households. On this basis, this paper builds a system of new rotating successive sampling survey on rural households by combining concrete national conditions with the theories of successive sampling survey, as involving rural households issue in the actual investigation workers guidance on the sampling method.In summary, the research on application and theory of split panel and rotating panel not only complements the successive sampling theory, but also play an important role in establishing the system of rural households sampling survey suitable for management system in China. First, the problems of rotating panel, panel samples and split panel are discussed deeply. And the research methods and results not only contribute to compare advantages and disadvantages of the four kinds of samples, but also make for the optimal choice of samples.Secondly, from the new perspective, the first form split panel is design optimally, and the optimization method is designed to solve the optimal design, which offer reliable theory basis for using the first form split panel. Thirdly, the estimation method using one half rotating panel is researched deeply, and the results show that the empirical research using one half rotating panel can offer reliable estimation results. At last, the theories of rotating panel are applied to sampling survey of China’s rural households, which is instructive to improve the system of rotating sampling survey of China’s rural households, and strongly fueled the implementation on sampling survey of China’s rural households. So it has theoretical and practical significance to research split panel and rotating panel.