Research On Estimation Of Census Net Error And Its Components

Due to the influence of population flow,non response and other factors,the total number of national population registration provided by the census inevitably deviates from the real total number of population.Therefore,census quality assessment has become an important part of census work.The main contents of this paper are the estimation methods of census net error,census under count and over count.The research goal of this paper is to explore the three system estimation method of census net error and the rate estimation method of census omission from the perspective of "the UK program".First of all,the basic theory,construction method and technical details of the British census are studied and interpreted in-depth,mainly from the four aspects of ONC project,CCS survey,estimation method,the difference between the "UK program" and "US program";secondly,the technical details of estimating the real population based on the three system estimates under the British thinking are explored,and at the same time,the general estimation is made This part mainly explains the division of the post event layer,the choice of the model,how to estimate the population without registration in three surveys under different models,and the structure of the estimation of the rate of the missing population.The innovation of this paper is the division of the post event layer and the estimation of the rate of the missing population.Finally,the sample survey of 1% of the national population in 2015 is carried out Based on the data,this paper uses the method of random simulation to verify the robustness of this research method,and demonstrates the construction process of this research method,providing guidance for relevant personnel,and providing valuable reproducible experience and method support for the seventh national census of China in 2020.The results of numerical simulation show that the net error of census estimated by the three system estimators under the "UK program" is 4298,the undercount is 3683.The results of random simulation show that the deviation,mean square error and mean variance of the three system estimators under the "UK program" are the smallest,.Therefore,it is suggested that China should consider using the three system estimators under the "UK program" to estimate the total population in the census year in the future census quality assessment.