femiguez/nlraa

psim = 3 for nlme - requires testing

femiguez opened this issue · 2 comments

Someday when I have a good excuse I will look into adding the option of full simulation in nlme from fixed, random and residual. This will probably not happen in the near future...

I have implemented this, but I'm a bit unsure about how well it works. Somewhat unexpectedly the total variance when I use psim = 3 is lower than when I use psim = 2. I really do not know the explanation for this.

I could also implement resampling sigma. Perhaps this would be psim = 4???

I will close this until I actually find a problem. It seems to be working.