大数据下定量敏感问题的调查方法  

An Investigation Method for Quantitative Sensitive Problems Under Big Data

作　　者：梁敏[1] Liang Min(School of Applied Mathematics,Beijing Normal University(Zhuhai),Zhuhai Guangdong 519087,China)

机构地区：[1]北京师范大学珠海分校应用数学学院,广东珠海519087

出　　处：《统计与决策》2021年第14期20-23,共4页Statistics & Decision

基　　金：珠海市哲学社会科学规划课题(2019ZC114);北京师范大学珠海分校教师科研能力促进计划项目。

摘　　要：目前定量敏感问题所用调查方法需使用随机化装置从而需要在现场实施。为此,文章充分利用大数据资源作为辅助信息,将分层抽样方法与不需要随机化装置的技术结合,实施精度最高的最优分配,方便大规模调查,不需要调查者亲临现场,适合计算机辅助调查,顺应大数据时代的要求。给出了相应的无偏估计量、推算出估计量的方差和方差的估计量。At present, the investigation methods used for quantitative sensitive problems need to use randomization devices and thus need to be implemented in the field. To this end, the paper makes full use of big data resources as auxiliary information to combine stratified sampling method with technology that does not require randomization devices, and implement the optimal allocation with the highest precision, which is convenient for large-scale investigation without the need for investigators to be present at the site, suitable for computer-aided investigation, and conforming to the requirements of the big data era. Finally, the paper gives the corresponding unbiased estimator, and calculates the variance of the estimator and the estimator of the variance.

关 键 词：敏感性问题 非随机化回答技术 大数据 

分 类 号：C811[社会学—统计学]