In hindsight... [Regulatives / Guidelines]

posted by ElMaestro  – Denmark, 2011-11-03 22:43 (4977 d 23:40 ago) – Posting: # 7599
Views: 9,182

Hi HS,

2 comments.

1:

ElMaestrosMixedModel <- lme(y1 ~ 0 + seq1 + per1 + trt1, random = ~1|sub1)

I am not sure why ElMaestro used a mixed model here; doesn't this specification just give the same as a linear model with seq1, per1, trt1, and sub1 as fixed factors :confused:?

2:
I think we need to take degrees of freedom into consideration too. After all, the critical value of the t-dist (and thus the width of the CI which is the ultimate indicator) depends heavily on it. I think the df's differ between the two scenarios. I hope this does not lead into a discussion about how df's are calculated for a mixed model (R will not do Satterthwaite etc.), a discussion where I have absolutely nothing to contribute.


I am sorry to question my own sanity.

Pass or fail!
ElMaestro

Complete thread:

UA Flag
Activity
 Admin contact
23,424 posts in 4,927 threads, 1,674 registered users;
39 visitors (0 registered, 39 guests [including 12 identified bots]).
Forum time: 23:24 CEST (Europe/Vienna)

Complex, statistically improbable things are by their nature
more difficult to explain than
simple, statistically probable things.    Richard Dawkins

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