Bioequivalence and Bioavailability Forum

Main page Policy/Terms of Use Abbreviations Latest Posts

 Log-in |  Register |  Search

Back to the forum  Query: 2017-10-19 00:17 CEST (UTC+2h)
 
Mohamed Yehia
Junior

Egypt,
2017-07-29 22:10

Posting: # 17626
Views: 596
 

 Confidence Interval For Tmax [Nonparametrics]

Dear All,

I want to know whether we can use transformed data on Tmax to check for CI?. AS it was stated previously that if we will take Tmax into statistical considerations we will take untransformed tmax not transformed ones.


Edit: Category changed; see also this post #1. [Helmut]
ElMaestro
Hero

Denmark,
2017-07-29 23:30

@ Mohamed Yehia
Posting: # 17628
Views: 509
 

 Confidence Interval For Tmax

Hi M.Y.,

» I want to know whether we can use transformed data on Tmax to check for CI?. AS it was stated previously that if we will take Tmax into statistical considerations we will take untransformed tmax not transformed ones.

CI's on Tmax are not much in use anymore. Don't bother.
I mainly see them popping up nowadays when applicants find themselves in desperate situations with deficiency letters and imminent rejection fears. That's when they try to generate all sorts of supportive funky BE results (which were never stipulated in any protocol or SAP) in the hope that quantity of tests outweighs the quality of a few well-chosen tests.:-):-):-)

I could be wrong, but…


Best regards,
ElMaestro

- since June 2017 having an affair with the bootstrap.
Mohamed Yehia
Junior

Egypt,
2017-07-29 23:38

@ ElMaestro
Posting: # 17629
Views: 506
 

 Confidence Interval For Tmax

» CI's on Tmax are not much in use anymore. Don't bother.

Hi Elmaestro,

Yes, I know that. However, I found a SAS report that generate a Non parametric analysis test on Ln transformed Tmax; thats why I am wondering is that accepted or not.


Edit: Full quote removed. Please delete everything from the text of the original poster which is not necessary in understanding your answer; see also this post #5! [Helmut]
Helmut
Hero
Homepage
Vienna, Austria,
2017-07-30 14:41

@ Mohamed Yehia
Posting: # 17630
Views: 484
 

 Confidence Interval For Tmax

Hi Mohamed,

first of all, please read this post and follow the forum’s policy (you confirmed with the registration that you have read and accepted it)! I had to edit all of your posts… :-(

» […] However, I found a SAS report that generate a Non parametric analysis test on Ln transformed Tmax;

tmax comes from an ordinal scale and therefore, follows a discrete distribution (dependent on the sampling schedule). Hence, a nonparametric analysis is correct in principle. However, transformations are only allowed for data coming from a ratio scale (continuous with a true zero point).
Therefore, the report you found is crap. SAS is a fine piece of software but only as good as the user. This one probably was not qualified for his/her job.

Common language will give you a hint whether an additive model (raw data) or a multiplicative model (log-transformed data) should be used. Every time you say “X is Y% of Z” or “X is Y% lower than Z” you should apply a log-transformation (e.g., Cmax). If you say “the difference of X to Z is Y” you should use the raw data (e.g., tmax).

[image]All the best,
Helmut Schütz 
[image]

The quality of responses received is directly proportional to the quality of the question asked. ☼
Science Quotes
Back to the forum Activity
 Thread view
Bioequivalence and Bioavailability Forum | Admin contact
17,394 Posts in 3,725 Threads, 1,072 registered users;
29 users online (0 registered, 29 guests).

No investigation can strictly be called scientific
unless it admits of mathematical demonstration.    Leonardo da Vinci

The BIOEQUIVALENCE / BIOAVAILABILITY FORUM is hosted by
BEBAC Ing. Helmut Schütz
XHTML/CSS RSS Feed