Multiple Imputation in Multivariate Problems When the Imputation and Analysis Models Differ

By Joseph L. Schafer. 2003

205 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/990c7457ff255ca023a6815c1e69eba0ca215698

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Statistics and Probability

Citations in the corpus

Multiple Imputation of Missing Data for Multilevel Models (2018, unknown): “Multiple Imputation for Multilevel Models A crucial point in the application of MI to multilevel data is that the imputation model not only includes all relevant variables, but also that it “matches” the model of interest (i.e., the substantive analysis model; see Meng, 1994; Schafer, 2003).”
Flexible Imputation of Missing Data (2012, background): “Another danger occurs if the imputation model is uncongenial (Meng, 1994; Schafer, 2003).”
Missing data techniques for multilevel data: implications of model misspecification (2011, background): “[29]), but we are aware of only one published study to date that has documented the relative quality of estimates under conditions of normal model misspecification with multilevel data [40].”
Multiple imputation as a flexible tool for missing data handling in clinical research. (2017, background): “…there is no theoretical reason to expect differences between maximum likelihood estimation and multiple imputation (Gelman et al., 2014; Meng, 1994; Schafer, 2003), and empirical studies suggest that the twomethods usually yield similar estimates and standard errors (Collins, Schafer, & Kam,…”
Multiple Imputation by Chained Equations: What is it and how does it work? (2011, methodology): “The second aspect of having an imputation model that is more general then the analysis model is including additional (“auxiliary”) variables in the imputation process – variables that are not going to be used in the analysis but that can improve the imputations (Collins et al., 2001; Schafer, 2003).”
Random Effects Models for Longitudinal Data (2010, methodology): “An alternative mode of analysis, generally overlooked but proposed by Schafer (2003), would consist in multiply imputing the missing outcomes using a parametric model, e.g., of a random-effects or conditional type, followed by conventional GEE and conventional multiple-imputation inference on the…”
International Large-Scale Assessment Data (2010, methodology): “This is analogous to a mismatch between the imputer’s model and the analyst’s model in the multiple imputation literature (Schafer, 2003).”
Combining Multiple Imputation and Inverse-Probability Weighting (2012, background): “Schafer (2003) comments that “although we may find it difficult to prove good performance for [MI using a nonmaximum likelihood estimator], that does not imply that good performance will not be seen in practice.”
Inverse Probability Weighting with Missing Predictors of Treatment Assignment or Missingness (2014, unknown): “Schafer (2003) comments that “although we may find it difficult to prove good performance for [the Rubin’s Rules variance estimator when not using the MLE], that does not imply that good performance will not be seen in practice.”
Multiple imputation using chained equations for missing data in TIMSS: a case study (2013, unknown): “This is not the case for individuals with more modest incomes.c Careful consideration of the missing-data mechanism is important because different types of missing data require different treatments (Allison, 2000; Schafer, 2003).”
Clinical reappraisal of the Composite International Diagnostic Interview Screening Scales (CIDI-SC) in the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS) (2013, methodology): “The first phase of MI develops prediction equations based on any of several different complex search methods (Schafer, 2003; White et al., 2011) to estimate multivariate associations of predictors with the variables to be imputed in the subset of respondents with complete data and to use those equat”
Multiple imputation of covariates by fully conditional specification: Accommodating the substantive model (2012, unknown): “9 .”