多重填补的方法及其统计推断原理

目的 描述数据缺失的特征和数据缺失模式，对Rubin最早提出的多重填补(multiple imputation,MI)的基本概念、填补和分析缺失数据的方法、综合统计推断进行了探讨，分析了MI的特点、局限性以及应用MI方法处理不完整数据集时需要注意的地方。方法 通过计算机模拟，用MI方法将每一个缺失值用一系列可能的值填补，然后使用常规的、针对完全数据集的统计方法对多重填补后得到的若干数据集进行分析，并把所得的结果进行综合。结果 多重填补值显示出了缺失数据的不确定性，使得已有数据得到了充分利用，从而对总体参数做出了更为准确的估计。结论 MI方法为处理存在缺失值的数据集提供了有用的策略，并且适用于多种数据缺失的场合。

Methods of Multiple Imputation and Related Inference Theory

Objective To describe the basic concepts,methods for imputing and analyzing missing data, combining inference of multiple imputation (MI) technique which was originally proposed by Rubin, and discuss limitations and cautions to consider when using MI as an analytic strategy for incomplete data settings. Methods By computer simulation, MI replaces each missing value with a set a plausible values. The multiply imputed data sets are then analyzed by using standard procedures for complete data and combining the results from these analyses. Results The multiply imputed data generated by MI procedure represent the uncertainty about the missing value and help us use data available to estimate population parameter more effectively. Conclusion MI provides a useful strategy for dealing with data sets with missing values and is broadly applicable to a variety of missing data settings.