I like to check that covariances among imputed variables are preserved as well as covariances between imputed variables and those variables not requiring imputation. With a large number of variables, this is generally an impossible goal, but one can at least check covariances among the most important analytic variables. Some statisticians like to check that marginal means are preserved, but I believe that this is a mistaken measure of quality. In fact, I hope that the marginal means are changed. Otherwise, there would appear to have been little point in undertaking the imputation.
David Judkins Senior Statistician Westat 1650 Research Boulevard Rockville, MD 20854 (301) 315-5970 [email protected] -----Original Message----- From: [email protected] [mailto:[email protected]] Sent: Tuesday, April 13, 2004 4:53 AM To: [email protected] Subject: [Impute] Imputation quality Dear Colleges! I am going to tailor the imputation procedure to one of the surveys. Therefore I am looking for the guidelines how to asses the quality of imputation procedure in a given case. Do you know any paper concerning this matter? Thank you for your help! Andrzej _______________________________________________ Impute mailing list [email protected] http://lists.utsouthwestern.edu/mailman/listinfo/impute
