Freedom for replicate bears [R for BE/BA]

posted by ElMaestro  – Denmark, 2011-01-27 21:50 (4011 d 23:08 ago) – Posting: # 6502
Views: 62,706

Dear dlabes and others,

» After reading again tons of Websites (seems nobody evaluates x-over studies with R) I came up with the following attempt in R:
»   model2 <- lme(y~ drug + prd + seq,
»        random= ~ tmt-1|subj,
»        #different within variabilities
»        weights=varIdent(form = ~ 1 | drug),
»        data=BlaBla, method="REML")

I am starting to believe lme does in some cases not really reach the goal, so I switched to lmer (package lme4). Syntactically something like:
ElModel=lmer(logCmax ~ per+seq+trt + (trt-1|subj))

-which achieves the same.
It will occasionally -depending on the dataset- actually fit the data better (that is, to a higher likelihood and thus with better sigma estimates) than lme can.

And also:
Who's got a 3-period dataset (ref. replicated, but not test-replicated) with a known sWR that I can look at?

Mucho gracias

Pass or fail!
ElMaestro

Complete thread:

Activity
 Admin contact
21,834 posts in 4,569 threads, 1,554 registered users;
online 12 (0 registered, 12 guests [including 3 identified bots]).
Forum time: Friday 20:59 CET (Europe/Vienna)

No problem can stand the assault of sustained thinking.    Voltaire

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