统计研究 ›› 2021, Vol. 38 ›› Issue (6): 116-127.doi: 10.19343/j.cnki.11-1302/c.2021.06.009

• • 上一篇    下一篇

规模以下工业抽样设计研究

万舒晨   

  • 出版日期:2021-06-25 发布日期:2021-06-25

Research on the Sampling Design of Industrial Surveys below the Designated Size

Wan Shuchen   

  • Online:2021-06-25 Published:2021-06-25

摘要: 为推动规模以下工业抽样调查工作以及解决当前调查面临的有关问题,本文对抽样设计进行了改进研究。首先,本文对规模以下工业抽样设计演变过程进行系统梳理,总结了现行抽样设计充分利用双重抽样框设计和综合运用三种抽样方法的特点。其次,针对园区层企业密度高的特点,探索结合园区因素改进地域抽样设计,对园区层和非园区层分别抽样,解决调查中面临的非抽样误差问题,并调整辅助变量使其与核心指标相关性均较高,确保抽样推断精度,有效提高抽样调查效率。并以我国东 部某省为例进行实证模拟得到结合园区因素抽样设计对调查工作改进的结论。再次,针对我国各级政府管理需要以及局队业务分工优化调整情况,介绍了规模以下工业样本追加理论和实证应用的主要研究成果。最后,在大数据时代数据来源广泛的背景下,本文在多重抽样框设计以及利用辅助变量提升样本轮换推断精度方面提出了进一步改进抽样设计的思路。

关键词: 规模以下工业, 抽样设计, 工业园区, 样本追加, 大数据

Abstract: In order to promote an industrial sampling survey below the designated size and to solve the current problems in the survey, this paper studies the improvement of the sampling design. Firstly, this paper systematically recalls the evolution process of the sampling design of industrial survey below the designated size and summarizes the characteristics of the current sampling design which uses double sampling frame design and three sampling methods. Secondly, considering the characteristic of a high density of enterprises in the industrial park stratum, this paper improves the current sampling design to sample in the park stratum and non- park stratum separately, solving the problem of non-sampling errors. This paper also adjusts the auxiliary variable to make it have high correlations with the core variables, to ensure the sampling accuracy and improve the efficiency of sampling surveys. Taking a province in eastern China as an example, this paper uses the empirical simulation and gets the conclusion that the sampling design considering industrial parks improves the survey. Thirdly, according to the management of governments at all levels and the business adjustment between statistical bureaus and survey offices, this paper introduces the main research results of the theory and empirical application for the sampling complement of industrial surveys below the designated size. Lastly, in view of the wide data sources in the context of big data, this paper proposes the idea of further improvement of sampling design in the aspects of multiple sampling frame design and using auxiliary variables to improve the accuracy of sampling rotation.

Key words: Industries below the Designated Size, Sampling Design, Industrial Park, Sampling Complement, Big Data