For the first question, I have a hunch that it might be due to a change in the denominator degrees of freedom as you have moved farm from a G side effect to an R side effect. Also, consider that the results for the F value do change when you go from an all G side parameterization to a G and R parameterization. Walt Stroup's text covers what could be happening with your situation.   The second part about why the number of events and trials differ so greatly between the two approaches is harder to answer. I am not sure how your data are structured, but I think that may be part of what is happening. For things like this, I have always trusted what @jiltao has said, and @Mike_N  has made some great comments recently.   SteveDenham ... View more

Your code looks like it may have been truncated as there's no model statement, and it looks like you have a colon instead of semi colon in the class statement. try adding param=ref to the class statement. This tells SAS to use Referential coding rather than the default of Effect coding for the categorical variables. SAS FAQ: In PROC LOGISTIC why aren't the coefficients consistent with the odds ratios? proc logistic data = ....; class treatment (ref=control) day (ref=0)/Param=REF; sick_animals/animals_per_cage= treatment day/cl; oddsratio treatment/diff=ref; oddsratio day/diff=ref; run; ... View more

This looks good--some minor typos (oddsration should be oddsratio, subjet should be subject).  I would change the type=un to type=chol, and would add method=quad or method=laplace to the proc statement.  To get: Proc glimmix method=quad data= class animal cohort farm x y: model y= x/dis=bin link=logit oddsratio; random intercept/subject=farm; random intercept/subject=animal(farm) type=un; run; I hope this helps. Steve Denham ... View more