jabriges
☆    

US,
2020-06-24 19:46

Posting: # 21577
Views: 900
 

 random effects in BEAR (lme_lm.mod.r) [R for BE/BA]

I have a question about getting subject level random effect. The call of lme() in lme_lm.mod.r uses

random=~drug - 1|subj

I'm not seeing how this gives you a random effect for each subject. I understand the -1 removes the intercept, but doesn't the inclusion of drug there make the random effect differ by formulation, and hence period (for the same subject)?
Any insight appreciated.

-Jay
ElMaestro
★★★

Belgium?,
2020-06-25 09:40

@ jabriges
Posting: # 21578
Views: 630
 

 random effects in BEAR (lme_lm.mod.r)

Hi all,

» random=~drug - 1|subj


I do not offer an answer to the question at all, only I wish to say that I have no ideas, generally, how syntactically R works for lme and lmer when specifying random effects.
All the stuff that involves "~", "|" for random effects is entirely obscure to me, and I tried to read the various online resources for it, and help files for packages. I even ordered the book by Pinheiro & Bates to learn from it, but that book assumes that the reader knows everything there is to about mixed models and R syntax before you open it on page 1. It is also not clear to me how covariance matrix structures are actually set using the R functions.

So, I think the question you asked jabriges is also highly relevant for me in a very general sense. I hope someone will chime in an educate me about the syntax.

I could be wrong, but...

Best regards,
ElMaestro

"Pass or fail" (D. Potvin et al., 2008)
jabriges
☆    

US,
2020-06-25 18:12

@ ElMaestro
Posting: # 21579
Views: 605
 

 R formulas

Hi, Maestro.

» So, I think the question you asked jabriges is also highly relevant for me in a very general sense. I hope someone will chime in an educate me about the syntax.

There are places that you can read about R's formula syntax in general, like here.

My question is specifically about the formula for random (see my OP) which seems to make the random coefficient a slope on the drug term, rather than a random intercept term (grouped by subject, of course), which was my understanding of the intended model.

-Jay
Activity
 Admin contact
20,793 posts in 4,353 threads, 1,445 registered users;
online 22 (0 registered, 22 guests [including 11 identified bots]).
Forum time: 19:33 CEST (Europe/Vienna)

In God we trust;
all others must bring data.    W. Edwards Deming

The Bioequivalence and Bioavailability Forum is hosted by
BEBAC Ing. Helmut Schütz
HTML5