Hi!!!!
I'm actually used the glimmix procedure to analyze some dates related to the effect of 4 experimental diets over hock burn occurrence in broiler chickens; the hock burn are 5 categories (1 to 5). I'm used the following editor (I considered the random effect of pen or "recinto"):
proc glimmix data=score method=laplace oddsratio;
class tratamiento recinto animal;
model hockburn (descending)=tratamiento/solution dist=multinomial link=cumlogit;
random recinto(tratamiento);
estimate '4 para 0L01P' intercept 4 0 0 0 tratamiento 4 0 0 0/ ilink;
estimate '3 para 0L01P' intercept 0 4 0 0 tratamiento 4 0 0 0/ ilink;
estimate '2 para 0L01P' intercept 0 0 4 0 tratamiento 4 0 0 0/ ilink;
estimate '1 para 0L01P' intercept 0 0 0 4 tratamiento 4 0 0 0/ ilink;
run;
In the ouoput results, I have a table with means of stimates but I'm confused with them: some values are e.g, 1.35E-8!. I think that this aren't the cummulative percentages of ocurrence, How do I to calculate them?
Thanks for help me!