Bioequivalence and Bioavailability Forum

Dear Helmut,

let me raise this old but useful post
first of all,

❝ A small step for a man but a giant leap for mankind?

should go to Detlew, take a look at his proposal
this is pretty the same, I didn't find any differences except some other rules of arguments naming.
I tried it for some models and it works well until uncomplete data is given:

> test(pairs(emmeans(mod, ~ Formulation, mode = "satterth"),
+            reverse = TRUE), delta = log(0.8))
Error in (mth$objs[[1]])(object, trms, xlev, grid, ...) :
  Unable to estimate Satterthwaite parameters

Any ideas? should I contact Russ regarding that?
It works when the mode is asymptotic or containment