Partial credit [Design Issues]

posted by ElMaestro  – Denmark, 2016-01-22 15:58 (3306 d 03:44 ago) – Posting: # 15846
Views: 11,011

Hi Ioana,

❝ since I have the "privilege" to work sometimes with Kinetica, I obtained the new version 5.1 SP1 pack released by Thermo Fisher in 2014, where they fixed the statistical part for the unbalanced studies.


May I recommend to Format C:

❝ Compared to the previous version, they mention now at the end of analysis:

❝ - ANOVA analysis: Point estimate and confidence interval derived from the least squares estimate (marginal means). - for a balanced study

❝ and for incomplete block analysis: Point estimate and confidence interval derived from the arithmetic means.


❝ Perhaps I am wrong, but calculation based by arithmetic means cannot give bias in such studies?


Yes, this sounds odd. Arithmetic means would work for parallel trials, but not necessarily for crossovers. It could be a misunderstanding or a matter of wording. Try and look up in the documentation what is going on - how it is calculated in practice, and please keep us posted :-)

Pass or fail!
ElMaestro

Complete thread:

UA Flag
Activity
 Admin contact
23,376 posts in 4,912 threads, 1,662 registered users;
234 visitors (0 registered, 234 guests [including 11 identified bots]).
Forum time: 19:42 CET (Europe/Vienna)

There are sadistic scientists who hurry to hunt down errors
instead of establishing the truth.    Marie Curie

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