We are using multiple imputation more frequently to “fill in” missing data in clinical datasets. Multiple datasets are created, models run, and results pooled so conclusions can be drawn. We’ve put some improvements into Finalfit on GitHub to make it easier to use with the mice package. These will go to CRAN soon but not immediately. In addition, we also found articles that used inverse probability weighting (7, 21, 22) or the expectation-maximization (EM) algorithm to account for missing data. The degree of detail reported in the 12 papers utilizing multiple imputation was highly variable. Seven papers stated the variables used to impute missing data (9, 10, 13–16, 19. It provides a discussion on how to choose different ancillary variables in the imputation model in either a prospective or a retrospective (exploratory) manner. After missingdata have been multiply imputed and each of the imputed datasets analyzed using a method suitable for the complete‐data analysis, the results from multiple datasets need. Simulations show the completed data after the new imputation approach have the proper distribution, and the estimators based on the new imputation method outperform the traditiona. Statistical Assumptions for Multiple Imputation. The MI procedure assumes that the data are from a multivariate distribution and contain missing values that can occur for any of the variables. It also assumes that the data are from a multivariate normal distribution when either the regression method or the MCMC method is used.
We evaluated multiple imputation methods (that produce multiple completed datasets, ... Multiple Imputation for Missing Values in Homicide Incident Data: An Evaluation Using Unique Test Data - John M. Roberts, Aki Roberts, Tim Wadsworth, 2018.