方法对比

	Multiple Imputation	Fay-Herriot模型（小区域估计）	分层抽样
领域	统计学	调查方法论	调查方法论
方法族	Process / pipeline	Regression model	Process / pipeline
起源年份	1987	1979	1977
提出者	Donald B. Rubin	Robert Fay & Roger Herriot	William G. Cochran
类型	Missing-data handling procedure	Model-based survey estimator	Probability-based survey sampling design
开创性文献	Rubin, D.B. (1987). Multiple Imputation for Nonresponse in Surveys. Wiley.	Fay, R. E., & Herriot, R. A. (1979). Estimates of income for small places: An application of James-Stein procedures to census data. Journal of the American Statistical Association, 74(366), 269–277.	Cochran, W. G. (1977). Sampling Techniques (3rd ed.). Wiley. ISBN: 978-0-471-16240-7
别名	MICE, Multivariate Imputation by Chained Equations, Çoklu Atama (Multiple Imputation — MICE)	SAE, Model-Based Small Area Estimation, Area-Level Model, Küçük Alan Tahmini	Proportional Stratified Sampling, Optimal Allocation Sampling, Stratum-Based Sampling, Tabakalı Örnekleme
相关	1	2	2
摘要	Multiple Imputation (MI), formally introduced by Donald B. Rubin in 1987, is a principled statistical procedure for handling missing data. Rather than replacing each missing value once, MI fills the gaps m times — each time drawing plausible values from the posterior predictive distribution of the missing data — producing m complete datasets. Each dataset is analysed independently, and the results are combined into a single set of estimates using Rubin's pooling rules. The MICE variant (Multivariate Imputation by Chained Equations), popularised by van Buuren and Groothuis-Oudshoorn (2011), extends the approach to mixed variable types by imputing each variable in turn through a sequence of conditional regression models.	Small Area Estimation (SAE) refers to statistical techniques that produce reliable estimates for subpopulations — geographical regions, demographic groups, or administrative units — where direct survey samples are too sparse to yield acceptable precision. The Fay-Herriot model, introduced by Robert Fay and Roger Herriot in 1979, is the canonical area-level SAE model. It supplements weak direct survey estimates with auxiliary covariate information through an empirical Bayes or BLUP framework, substantially reducing mean squared error for small domains.	Stratified sampling is a probability sampling design in which the target population is partitioned into non-overlapping, exhaustive subgroups called strata, and independent probability samples are drawn within each stratum. Formalized by William G. Cochran in Sampling Techniques (1977), the method exploits known population structure to reduce variance and guarantee representativeness of all major subgroups, making it a cornerstone of large-scale survey research and official statistics.