Thanks Dale,

Importantly- this was the key bit of information I was missing "The MIXED procedure data sets are constructed in the same way as the data you show from the GLIMMIX procedure." I did review this document- but then was not sure how to adapt it for GLIMMIX.

Much appreciated, Mah-J

-----Original Message----- From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU]On Behalf Of Dale McLerran Sent: Thursday, February 16, 2006 1:22 PM To: SAS-L@LISTSERV.UGA.EDU Subject: Re: transposing glimmix output for mianalyze

Mah-J,

Because you are using the GLIMMIX procedure, then I can unequivocally state that you are running SAS version 9.1. And since you have version 9.1, then I can also tell you that you do not need to transpose the parameter estimates data set. When an experimental version of MIANALYZE was released in SAS version 8.1, you could not process directly in MIANALYZE results from procedures which produce a data set of parameter estimates such as those generated by the GLIMMIX procedure. But with the production release of MIANALYZE in SAS version 9.0, you can analyze data like those generated by GLIMMIX without the need for any intermediate processing.

See the example in the procedure MIANALYZE documentation where results from the MIXED procedure are passed through MIANALYZE. The MIXED procedure data sets are constructed in the same way as the data you show from the GLIMMIX procedure. No transpose is necessary for the data from the MIXED procedure. The example for the MIXED procedure can be found at http://support.sas.com/onlinedoc/913/getDoc/en/statug.hlp/mianalyze_sect22.h tm

I would note, too, that your MIANALYZE syntax is not syntax that is supported for the version of SAS that you are running. You are using experimental MIANALYZE syntax. The production version no longer employs the VAR statement. You will see this in the example pointed to above. If you replace "sect22" with "index" in the above URL, then you can read up-to-date documentation of the MIANALYZE procedure.

Dale

--- "mah-j@statworks.com" <msoobader@VERIZON.NET> wrote:

> Hi Nat, > > My apologies for not being any clearer- but I really don't know how > the data > should look when they are transposed- all I got from SAS technical > support > was that I needed to transpose the data. "If you PROC PRINT the > parameter > estimates dataset from GLIMMIX, it is not in the form that PROC > MIANALYZE > requires. You will need to transpose that dataset to get the > structure of > having separate variables for each parameter estimate." And I can't > find any > documentation on this. > > Both datasets printed below (COVB and PESTS) are output from GLIMMIX. > And > both need to be read by MIANALYZE. These contain the covariance > matrix of > the parameter estimates and the fixed effect parameters respectively. > > Hope this clarifies some. > Thank you, > Mah-J > > > Dsn=covb > > Obs _imputation_ Effect Row Col1 Col2 > Col3 Col4 Col5 Col6 > > 1 1 Intercept 1 > 0.000788 -0.00002 -0.00012 -0.00014 -0.00011 -0.00012 > 2 1 ageyr 2 -0.00002 > 2.288E-6 -3.13E-8 3.872E-6 8.221E-7 2.202E-6 > 3 1 male 3 -0.00012 -3.13E-8 > 0.000222 3.557E-6 4.503E-6 5.385E-6 > 4 1 hispanic 4 -0.00014 3.872E-6 > 3.557E-6 0.000483 0.000099 0.000115 > 5 1 nh_black 5 -0.00011 8.221E-7 > 4.503E-6 0.000099 0.000640 0.000091 > 6 1 nh_other 6 -0.00012 2.202E-6 > 5.385E-6 0.000115 0.000091 0.000820 > 7 2 Intercept 1 > 0.000789 -0.00002 -0.00012 -0.00014 -0.00011 -0.00012 > 8 2 ageyr 2 -0.00002 > 2.287E-6 -3.17E-8 3.875E-6 8.075E-7 2.217E-6 > 9 2 male 3 -0.00012 -3.17E-8 > 0.000222 3.443E-6 3.88E-6 5.239E-6 > 10 2 hispanic 4 -0.00014 3.875E-6 > 3.443E-6 0.000483 0.000099 0.000115 > 11 2 nh_black 5 -0.00011 8.075E-7 > 3.88E-6 0.000099 0.000640 0.000091 > 12 2 nh_other 6 -0.00012 2.217E-6 > 5.239E-6 0.000115 0.000091 0.000820 > > dsn=pests > > Obs _imputation_ Effect Estimate > StdErr > DF tValue Probt > > 1 1 Intercept 1.7100 > 0.02806 > 50 60.93 <.0001 > 2 1 ageyr -0.03770 > 0.001512 > 95733 -24.92 <.0001 > 3 1 male -0.1271 > 0.01491 > 95733 -8.53 <.0001 > 4 1 hispanic -1.1245 > 0.02197 > 95733 -51.17 <.0001 > 5 1 nh_black -0.7711 > 0.02530 > 95733 -30.48 <.0001 > 6 1 nh_other -0.4542 > 0.02863 > 95733 -15.87 <.0001 > 7 2 Intercept 1.7071 > 0.02810 > 50 60.76 <.0001 > 8 2 ageyr -0.03768 > 0.001512 > 95733 -24.92 <.0001 > 9 2 male -0.1246 > 0.01491 > 95733 -8.36 <.0001 > 10 2 hispanic -1.1197 > 0.02198 > 95733 -50.94 <.0001 > 11 2 nh_black -0.7671 > 0.02530 > 95733 -30.32 <.0001 > 12 2 nh_other -0.4490 > 0.02863 > 95733 -15.68 <.0001 > > > ods output parameterestimates=pests covparms=cp covb=covb ; > proc glimmix data=one; > class state; > model dvexvg (EVENT='1')=ageyr male hispanic nh_black nh_other/ > solution > covb link=logit dist=binomial oddsratios; > random int / sub=state; > random _residual_; > by _imputation_; > nloptions tech=nrridg; > run; > /* transformation ??*/ > proc mianalyze parms=tPESTS covb=tcovb ; > var intercept ageyr male ; > run; > > > > ----------------------------------------- > CONFIDENTIALITY NOTICE: This electronic message contains > information which may be legally confidential and/or privileged and > does not in any case represent a firm ENERGY COMMODITY bid or offer > relating thereto which binds the sender without an additional > express written confirmation to that effect. The information is > intended solely for the individual or entity named above and access > by anyone else is unauthorized. If you are not the intended > recipient, any disclosure, copying, distribution, or use of the > contents of this information is prohibited and may be unlawful. If > you have received this electronic transmission in error, please > reply immediately to the sender that you have received the message > in error, and delete it. Thank you. >

--------------------------------------- Dale McLerran Fred Hutchinson Cancer Research Center mailto: dmclerra@NO_SPAMfhcrc.org Ph: (206) 667-2926 Fax: (206) 667-5977 ---------------------------------------

__________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com