topic Missing Data Problem in Statistical Procedures https://communities.sas.com/t5/Statistical-Procedures/Missing-Data-Problem/m-p/162015#M8420 <HTML><HEAD></HEAD><BODY><P class="user-details">Hi folks </P><P></P><P>Given the fact that most imputation methods assume multivariate normality, does it make sense to impute missing values on a categorical variable let's say Gender and how can that be done? </P><P></P><P>Thanks</P></BODY></HTML> Wed, 24 Dec 2014 18:56:46 GMT ccasboy 2014-12-24T18:56:46Z Missing Data Problem https://communities.sas.com/t5/Statistical-Procedures/Missing-Data-Problem/m-p/162015#M8420 <HTML><HEAD></HEAD><BODY><P class="user-details">Hi folks </P><P></P><P>Given the fact that most imputation methods assume multivariate normality, does it make sense to impute missing values on a categorical variable let's say Gender and how can that be done? </P><P></P><P>Thanks</P></BODY></HTML> Wed, 24 Dec 2014 18:56:46 GMT https://communities.sas.com/t5/Statistical-Procedures/Missing-Data-Problem/m-p/162015#M8420 ccasboy 2014-12-24T18:56:46Z Re: Missing Data Problem https://communities.sas.com/t5/Statistical-Procedures/Missing-Data-Problem/m-p/162016#M8421 <HTML><HEAD></HEAD><BODY><P><A href="http://www.ats.ucla.edu/stat/sas/seminars/missing_data/part1.htm" title="http://www.ats.ucla.edu/stat/sas/seminars/missing_data/part1.htm">Multiple Imputation in SAS, Part 1</A></P></BODY></HTML> Wed, 24 Dec 2014 19:35:51 GMT https://communities.sas.com/t5/Statistical-Procedures/Missing-Data-Problem/m-p/162016#M8421 Reeza 2014-12-24T19:35:51Z Re: Missing Data Problem https://communities.sas.com/t5/Statistical-Procedures/Missing-Data-Problem/m-p/162017#M8422 <HTML><HEAD></HEAD><BODY><P>Though you could impute gender using a multiple imputation method that assumes normality, better approaches are available. The MI procedure includes fully conditional specification (FCS) multiple imputation using the FCS statement. The beauty of FCS multiple imputation is that it allows the user to employ logistic regression for imputing some variables and OLS regression for imputing others. The link below provides an example of using logistic regression to impute a nominal variable (species) and OLS regression to impute the other variables in the data set (length and width).</P><P></P><P><A href="http://support.sas.com/documentation/cdl/en/statug/66859/HTML/default/viewer.htm#statug_mi_examples08.htm" title="http://support.sas.com/documentation/cdl/en/statug/66859/HTML/default/viewer.htm#statug_mi_examples08.htm">SAS/STAT(R) 13.1 User's Guide</A></P><P></P><P>I believe that the FCS statement became production in SAS/STAT 13.1. So, you'll need a recent version of SAS to perform FCS multiple imputation.</P><P></P><P>Best,</P><P>Steve</P></BODY></HTML> Mon, 29 Dec 2014 05:26:41 GMT https://communities.sas.com/t5/Statistical-Procedures/Missing-Data-Problem/m-p/162017#M8422 StatsGeek 2014-12-29T05:26:41Z