Log in
Registeradjusted means (aka LSMs) [Study Performance]
posted by Helmut 
– Vienna, Austria, 2014-07-27 04:23 (3927 d 00:36 ago) – Posting: # 13318
Views: 18,072
HI Khaoula,
Yes, it’s wrong. Where did you find this formula? You would get a negative value for the root. Maybe Kinetica uses (also wrongly) CV% = 100√ℯMSE² – 1 which would give 4.72% – at least closer to your original 3.47% – no idea. It’s not my job to unveil all potential bugs in a software I neither seriously used nor validated myself…
Again, no idea. How did you calculate a GMR of 1.21? Your corrected log-transformed data gives a LSM for the reference of 6.1477 and for the test 6.3125. These values are derived exactly as I stated in my old post.
Mean of R in sequence RT is 6.0203 and in sequence TR is 6.2751. LSM = (xRT+xTR)/2 = 6.1477. ✔
Similarily we get the LSM of the test 6.3125. ✔
Backtransformed LSMs are ℯ6.1477 = 467.64 (R) and ℯ6.3125 = 551.40 (T). Therefore, the GMR is 551.40/467.64 = 1.1791, agreeing with what PHX/WNL reports.
PS: Sometimes in study reports the geometric means of treatments and results of the ANOVA are given in the same table (without the LSMs). Then – if the study was not balanced – people are confused, since the ratio of geometric means does not match the point estimate (calculated from the LSMs). Here the backtransformed geometric means are 471.64 (R) and 557.90 (T). Wrongly calculating the ratio 557.90/471.64 = 1.1829 ≠ 1.1791…
❝ the software […] take this rule: ![[image]](img/uploaded/image261.png)
(it's false)
Yes, it’s wrong. Where did you find this formula? You would get a negative value for the root. Maybe Kinetica uses (also wrongly) CV% = 100√ℯMSE² – 1 which would give 4.72% – at least closer to your original 3.47% – no idea. It’s not my job to unveil all potential bugs in a software I neither seriously used nor validated myself…
❝ but for GMR, I'm really confused, It's an unbalenced cross over so you told me to use the least squares (or adjusted) means instead I read this post
❝
❝ I have'nt found the same result (1,21), but 1,18 like kinetica v4... where is the problem?
Again, no idea. How did you calculate a GMR of 1.21? Your corrected log-transformed data gives a LSM for the reference of 6.1477 and for the test 6.3125. These values are derived exactly as I stated in my old post.
subj seq form per logCmax subj seq form per logCmax
────────────────────────────────────────────────────
1 RT R 1 6.6399 3 TR R 2 6.4505
2 RT R 1 5.7104 4 TR R 2 6.1485
7 RT R 1 5.7268 5 TR R 2 6.3439
8 RT R 1 5.8944 6 TR R 2 6.2634
9 RT R 1 5.4931 11 TR R 2 5.9162
10 RT R 1 6.0379 12 TR R 2 6.5582
14 RT R 1 6.6399 13 TR R 2 7.0733
15 RT? 16 TR R 2 5.4467
────────────────────────────────────────────────────
x 6.0203 x 6.2751subj seq form per logCmax subj seq form per logCmax
────────────────────────────────────────────────────
1 RT T 2 6.6134 3 TR T 1 6.6529
2 RT T 2 5.3753 4 TR T 1 6.4425
7 RT T 2 5.4806 5 TR T 1 6.3386
8 RT T 2 6.2748 6 TR T 1 6.3026
9 RT T 2 6.2500 11 TR T 1 6.5568
10 RT T 2 6.3491 12 TR T 1 6.5425
14 RT T 2 6.6134 13 TR T 1 7.2442
15 RT? 16 TR T 1 5.8260
────────────────────────────────────────────────────
x 6.1367 x 6.4883Mean of R in sequence RT is 6.0203 and in sequence TR is 6.2751. LSM = (xRT+xTR)/2 = 6.1477. ✔
Similarily we get the LSM of the test 6.3125. ✔
Backtransformed LSMs are ℯ6.1477 = 467.64 (R) and ℯ6.3125 = 551.40 (T). Therefore, the GMR is 551.40/467.64 = 1.1791, agreeing with what PHX/WNL reports.
PS: Sometimes in study reports the geometric means of treatments and results of the ANOVA are given in the same table (without the LSMs). Then – if the study was not balanced – people are confused, since the ratio of geometric means does not match the point estimate (calculated from the LSMs). Here the backtransformed geometric means are 471.64 (R) and 557.90 (T). Wrongly calculating the ratio 557.90/471.64 = 1.1829 ≠ 1.1791…
—
Dif-tor heh smusma 🖖🏼 Довге життя Україна!![[image]](https://static.bebac.at/pics/Blue_and_yellow_ribbon_UA.png)
Helmut Schütz
![[image]](https://static.bebac.at/img/CC by.png)
The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz
The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
Complete thread:
- randomization khaoula 2014-06-06 14:04 [Study Performance]
![[Mix]](https://static.bebac.at/img/mix.png "open in Mix view")
- Imbalanced cross-overs Helmut 2014-06-06 14:47
- Imbalanced cross-overs khaoula 2014-06-06 16:03
- CV 0.03%? Helmut 2014-06-07 13:52
- Imbalanced cross-overs jag009 2014-06-09 15:36
- Imbalanced cross-overs khaoula 2014-06-09 22:07
- Bag full of bugs Helmut 2014-06-10 00:37
- Randomisation? ElMaestro 2014-06-10 01:00
- Randomisation? Helmut 2014-06-10 14:53
- Randomisation? khaoula 2014-06-11 00:17
- Problems resolved? Helmut 2014-06-11 02:01
- Problems resolved? khaoula 2014-07-24 13:36
- corrected data in PHX/WNL Helmut 2014-07-27 01:37
- Problems resolved? khaoula 2014-07-24 13:36
- Problems resolved? Helmut 2014-06-11 02:01
- Randomisation? khaoula 2014-06-11 00:17
- Randomisation? Helmut 2014-06-10 14:53
- Randomisation? ElMaestro 2014-06-10 01:00
- Which version of Kinetica? Helmut 2014-06-11 19:51
- Which version of Kinetica? khaoula 2014-06-12 00:41
- Which version of Kinetica? khaoula 2014-07-25 00:49
- adjusted means (aka LSMs)Helmut 2014-07-27 02:23
- adjusted means (aka LSMs) khaoula 2014-08-25 21:41
- adjusted means (aka LSMs)Helmut 2014-07-27 02:23
- Bag full of bugs Helmut 2014-06-10 00:37
- Imbalanced cross-overs khaoula 2014-06-09 22:07
- Imbalanced cross-overs khaoula 2014-06-06 16:03
- Imbalanced cross-overs Helmut 2014-06-06 14:47
Ing. Helmut Schütz