Our new command midiagplots makes diagnostic plots for multiple imputations created by mi impute. The plots compare the distribution of the imputed values with that of the observed values so that problems with the imputation model can be corrected before the imputed data are analyzed. We include an example and suggest extensions to other diagnostics.
AbayomiK., GelmanA., and LevyM.2008. Diagnostics for multivariate imputations. Journal of the Royal Statistical Society, Series C57: 273–291.
2.
CarlinJ. B., GalatiJ. C., and RoystonP.2008. A new framework for managing and analyzing multiply imputed data in Stata. Stata Journal8: 49–67.
3.
CarlinJ. B., LiN., GreenwoodP., and CoffeyC.2003. Tools for analyzing multiple imputed datasets. Stata Journal3: 226–244.
4.
EndersC. K.2010. Applied Missing Data Analysis.New York: Guilford Press.
5.
GelmanA., KingG., and LiuC.1998. Not asked and not answered: Multiple imputation for multiple surveys. Journal of the American Statistical Association93: 846–857.
6.
LittleR. J. A., and RubinD. B.2002. Statistical Analysis with Missing Data. 2nd ed. Hoboken, NJ: Wiley.
RoystonP.2004. Multiple imputation of missing values. Stata Journal4: 227–241.
10.
RubinD. B.1987. Multiple Imputation for Nonresponse in Surveys.New York: Wiley.
11.
WhiteI. R., RoystonP., and WoodA. M.2011. Multiple imputation using chained equations: Issues and guidance for practice. Statistics in Medicine30: 377–399.
