Much ado about nothing: A comparison of missing data methods and software to fit incomplete data regression models.

摘要: Missing data are a recurring problem that can cause bias or lead to inefficient analyses. Statistical methods address missingness have been actively pursued in recent years, including imputation, likelihood, and weighting approaches. Each approach is more complicated when there many patterns of missing values, both categorical continuous random variables involved. Implementations routines incorporate observations with incomplete regression models now widely available. We review these the context motivating example from large health services research dataset. While still limitations current implementations, additional efforts required analyst, it feasible partially observed should be used practice.