proc glimmix SAS 9.3 and SAS 9.4

AndreyMyslivets — Fri, 15 Mar 2019 08:20:36 GMT

Hello everyone,

Could you please help me with following question?

Some time ago the study was conducted and analyzed in SAS 9.3 and one of the analyses was provided using the following script:

ods output Tests3 = ANOVA ;
proc glimmix data = cols1 ;
  class q1 y site usubjid;
  model  chis12 =  q1 y q1*y site chis /  dist=negbin offset = offs  ;
  random y  / subject=usubjid type=ar(1) solution ;
  lsmeans q1*y/ diff cl ; 
run;

At the moment the data with absolutely the same nature was collected and after running the SAS code on the new data, we have the following results:

Could you please help to understand why there are no p-values for each effect? What is the best option for checking the data in this situation and what conclusions we can make here?

P.S. if you see any mistakes, please highlight it.

Thanks!

Re: proc glimmix SAS 9.3 and SAS 9.4

PaigeMiller — Fri, 15 Mar 2019 12:52:55 GMT

You have zero degrees of freedom for the denominator, so no F-test is possible.

Now, why might you have zero degrees of freedom for the denominator? It depends on the design of the study, and how many data points you have and how much true replication you have. Those are things we can't answer without a lot more information from you.

topic Re: proc glimmix SAS 9.3 and SAS 9.4 in Statistical Procedures

proc glimmix SAS 9.3 and SAS 9.4

Re: proc glimmix SAS 9.3 and SAS 9.4