Log in
Register|
MaxPerry ☆ 2011-03-02 12:30 (5584 d 11:08 ago) Posting: # 6681 Views: 11,887 |
|
|
In the ANOVA model commonly used in bioequivalence testing (2-period crossover: factors Period, Treatment, Sequence, Subject nested within sequence, Residual), the significance of Sequence is tested sometimes using MS Error, and sometimes using MS Subject within Sequence. I have been unable to find an explanation of this, and would appreciate assistance. |
|
ElMaestro ★★★ Denmark, 2011-03-02 13:03 (5584 d 10:34 ago) @ MaxPerry Posting: # 6682 Views: 11,064 |
|
|
Hi MaxPerry, Seq is a between-subject factor. Therefore the meaningful F-test denominator thereby is the between-subject error. In the typical anova the residual error is a measure of intra-subject variability so this is is used for the within-factors. — Pass or fail! ElMaestro |
|
d_labes ★★★ Berlin, Germany, 2011-03-02 14:09 (5584 d 09:28 ago) @ MaxPerry Posting: # 6685 Views: 11,015 |
|
|
— Regards, Detlew |
.|
MaxPerry ☆ 2011-03-03 12:46 (5583 d 10:51 ago) @ d_labes Posting: # 6692 Views: 11,064 |
|
|
Thank you ElMaestro and d_labes. ElMaestro: This certainly seems rational. On the other hand, consider a simple ANOVA with multiple measurements of a particular feature on each subject in a group. The test of significance between subjects has to rely upon a within-subject error estimate. d_labes: In the thread you referenced, HS states ❝ sequences are a between subject effect and therefore need to be compared to the between subject residual error - which corresponds to ElMaestro's point. But the thread ultimately seems to settle on the explanation that the important distinction was whether subject effects were considered fixed or random. I'll read about this in a book and try to get my head around it. Regards Max |
|
ElMaestro ★★★ Denmark, 2011-03-04 10:42 (5582 d 12:56 ago) @ MaxPerry Posting: # 6702 Views: 11,064 |
|
|
Hi MaxPerry, ❝ ElMaestro: This certainly seems rational. On the other hand, consider a simple ANOVA with multiple measurements of a particular feature on each subject in a group. The test of significance between subjects has to rely upon a within-subject error estimate. To me this certainly sounds like a mixed model and thus not a simple ANOVA because the feature response is correlated within subjects. Proc Mixed, not Glm. ❝ But the thread ultimately seems to settle on the explanation that the important distinction was whether subject effects were considered fixed or random. I'll read about this in a book and try to get my head around it. Uh oh - the Bogus Statement Warning Light is now flashing ("BSWL" for future reference). In GLM subjects are treated as fixed; SAS then has defined the bogus statement (search the forum to dig deeper into this term) which pretends to involve setting subjects as random, but the actual fit is still done with subjects as fixed (yeah, it has to be so because it is a GLM). My proposed bottom line: In a GLM (that is to say, in a 2,2,2-BE study) subjects are fixed; post-fit bogus gymnastics ascertains p-value with the between-subject variability in the F-test denominator for Seq. In a study with replication, within-subject responses to the same treatment are (can be) correlated and therefore a mixed model is used. — Pass or fail! ElMaestro |
Ing. Helmut Schütz