Inconsistencies between parminfo and covB parameters

lyton80 — Thu, 31 Aug 2017 16:53:21 GMT

I am using proc genmod and proc mianalyze. There is a discrepancy between the number of parameters in the genmod (with noscale option) and Cov parameter list.

proc genmod data=Edeathm2;
class issstg mmtg;
model pdgrp1 = mmtg issstg ECOGG cyto2rg ldhg/dist=bin noscale lrci covb;
by _Imputation_;
ods output ParameterEstimates=lgparms CovB=lgcovb parminfo= parmdata;
run;

proc mianalyze parms(classvar= level)= lgparms covb=lgcovb parminfo=parmdata;
class issstg mmtg;
modeleffects Intercept mmtg issstg ECOGG cyto2rg ldhg;
ods output ParameterEstimates=est1;
run;

I am returning this error: "ERROR: The parameter Prm4 in the PARMINFO= data set is not in the COVB= data set".

when I run proc mianalyze. Thanks in advance for the assistance.

Re: Inconsistencies between parminfo and covB parameters

SAS_Rob — Fri, 01 Sep 2017 18:24:19 GMT

All MIANALYZE needs is the PARMS= data set. The only time you need the other two data sets is when you have a TEST statement or use the MULT option, neither of which are available with a CLASS statement.

This should work:

proc mianalyze parms(classvar= level)= lgparms;
class issstg mmtg;
modeleffects Intercept mmtg issstg ECOGG cyto2rg ldhg;
ods output ParameterEstimates=est1;
run;

That being said,

topic Re: Inconsistencies between parminfo and covB parameters in Statistical Procedures

Inconsistencies between parminfo and covB parameters

Re: Inconsistencies between parminfo and covB parameters