05-31-2013 12:11 AM
I have created five datasets after multiple imputation however, when I use
proc mianalyze parms=gmparms covb=gmcovb parminfo=gmpinfo;
I get the following error
ERROR: The parameter Prm3 in the PARMINFO= data set is not in the COVB= data set.
I understand it has something to do with my categorical variables..Prm 2 Prm 3 represent gender, where prm3 is the reference category similarly Prm4,5, represent SES & Prm6 is the reference category..6 .But I am not sure how to fix it. I tried to include class statement but in vain. Any help would be appreciated.
05-31-2013 08:27 AM
Check out the discussion:
and somewhere down near the bottom, we finally figured out most of it. It still looks like the global tests were difficult to construct.
05-31-2013 11:28 AM
Thank you for providing me the link to the discussion. It is very helpful. I have read the documentation avaialble at
However, I am still stuck at the part where one has to combine the results of PROC GENMOD (for all the five imputations) which are stored in an output file now. Mostly as you pointed out in a prior discussion because I have many CLASS variables. If I try the code below I still get error
proc mianalyze parms(classvar=full)=gmparms;
class GENDER grage gr_revenu2 SM02 CARDIAC_RELATED_COND PHYSICIAN NAM PATIENT NEUROTIC_DISORDER;
modeleffects GENDER grage gr_revenu2 SM02 CARDIAC_RELATED_COND NEUROTIC_DISORDER
centred_AC centred_AQ centred_BM centred_SE ;
I get the following error which doesnt make sense because this variable is very much in the data set
ERROR: Variable GENDER is not in the PARMS= data set.
I would really appreciate any help.
06-03-2013 09:31 AM
I'm sorry Tasneem, but this is one I don't have an answer for. I suggest contacting Tech Support and opening a ticket with them. If you do get this solved, please post back here so that we can know what to do.
06-03-2013 01:47 PM
Thanks for your help. I have mailed the cite manager for SAS at our University to open a technical support ticket. Will keep u guys posted.
By the way do you open techincal support tickets through using the site number for SAS licnecne or is there a quicker way around.
06-04-2013 04:22 PM
I was able to figure out the solution to the problem, after getting help from SAS tech support.
The correct specification for the CLASSVAR= option from GENMOD is LEVEL.Hence the correct statement would be
proc mianalyze parms(classvar=level)=gmparms edf=218;
Hope it helps others with similar issues.
Thanks again for the help.
Need further help from the community? Please ask a new question.