基于“三新”经济动态总体的事后分层抽样方法  

Post Stratified Sampling Method Based on"Three New"Economic Dynamic Population

作　　者：张维群 成鹏东 杨静宜 ZHANG Wei-qun;CHENG Peng-dong;YANG Jing-yi(School of statistics,Xi'an University of Finance and Economics,Xi'an 710100,China;School of Statistics and Data Science,Nanjing Audit University,Nanjing 210000,China)

机构地区：[1]西安财经大学统计学院,陕西西安710100 [2]南京审计大学统计与数据科学学院,江苏南京210000

出　　处：《数理统计与管理》2024年第6期951-961,共11页Journal of Applied Statistics and Management

基　　金：国家社科基金重大项目(21&ZD147)。

摘　　要：“三新企业”是新经济企业的重要组成部分,科学地抽样推断“三新企业”规模及水平有利于准确及时地掌握新经济发展的现状,对于“三新企业”的抽样调查也是当前经济统计学界关注的主要问题。文章基于固定抽样框无法反映“三新企业”抽样调查中总体容量及结构频繁变动的特点,设计了基于“三新”经济动态总体的事后分层抽样方法,首先通过多重连续抽样,实现对总体容量变动的估计,以及对动态抽样框的修正和事后分层处理,然后依据复合抽样技术,构建基于“三新”经济动态总体相关参数的复合估计量,并对估计量的优良性进行讨论。模拟结果显示,基于“三新”经济动态总体的事后分层抽样方法能够实现目标变量的有效估计,构造的总体参数估计量具有无偏性和有效性。"Three new enterprises"are an important part of new economic enterprises.Scientifically sampling and inferring the scale and level of"three new enterprises"is conducive to accurately and timely grasp the current situation of new economic development.The sampling survey of"three new enterprises"is also the main concern of the current field of economic statistics.Based on the fact that the permanent sampling frame can not refect the frequent changes of the population size and stratification structure in the sampling survey of"three new enterprises",this paper designs a composite sampling method based on the post-stratification of the dynamic population.Through multiple continuous sampling,it realizes the estimation of the population size change,the correction of the dynamic sampling frame and the post-stratification processing.Then,according to the multisampling technology,it designs a composite estimator based on the relevant parameters of the dynamic population,And in this article,sampling error is defined by theoretic derivation.the sampling error.The laboratory results show that the composite sampling method based on dynamic population post-stratification can effectively estimate the object variables,and according to this method,the estimator of population parameters is unbiased and eficient estimator.

关 键 词：动态总体 事后分层 多重抽样 复合估计量 

分 类 号：C811[社会学—统计学] O212[理学—概率论与数理统计]