imbalanced design? [Software]
Dear Eva!
One example in SAS…
… and WinNonlin 5.2
Obviously sequence effects are identical (p 0.187). Maybe you had some drop-outs in your study, which resulted in imbalanced sequences. In such a case SAS’ SS Type III and WinNonlin’s Partial Sum of Squares may differ (although they ‘coincide in most situations’). SAS SS III is a strange cup of tea anyhow…
It would be nice, if you would come up with a simple dataset (12 subjects?), where you find differences. It should be possible to track it down!
One example in SAS…
SUM OF
SQUARES MEAN
EFFECT TYPE I DF SQUARE F-VALUE P-VALUE
----------------------------------------------------
TREATMENT 0.956 1 0.956 7.961 0.007
SUBJ. WITHIN SEQ. 17.802 46 0.387 3.224 <.001
PERIOD 0.100 1 0.100 0.835 0.366
SEQUENCE 0.694 1 0.694 1.794 0.187
ERROR 5.521 46 0.120
… and WinNonlin 5.2
Partial Sum of Squares
Hypothesis DF SS MS F_stat P_value
-----------------------------------------------------
SEQUENCE 1 0.694182 0.694182 1.79377 0.1870
SEQUENCE*SUBJ. 46 17.8019 0.386997 3.22418 0.0001
TREATMENT 1 0.955579 0.955579 7.9612 0.0070
PERIOD 1 0.100213 0.100213 0.834903 0.3656
Error 46 5.52136 0.12003
Obviously sequence effects are identical (p 0.187). Maybe you had some drop-outs in your study, which resulted in imbalanced sequences. In such a case SAS’ SS Type III and WinNonlin’s Partial Sum of Squares may differ (although they ‘coincide in most situations’). SAS SS III is a strange cup of tea anyhow…
It would be nice, if you would come up with a simple dataset (12 subjects?), where you find differences. It should be possible to track it down!
—
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
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:
- SAS vs. WinNonlin: different sequence effect results Eva 2008-05-06 14:04 [Software]
- SAS vs. WinNonlin: different sequence effect results Ohlbe 2008-05-06 15:56
- imbalanced design?Helmut 2008-05-06 16:13
- Sample data & results Eva 2008-05-06 17:42
- Error factor Ohlbe 2008-05-06 19:25
- Error tests between/within Helmut 2008-05-06 21:33
- Error tests between/within Helmut 2008-05-07 18:18
- The power to know d_labes 2008-05-08 11:16
- The power to know Helmut 2008-05-08 17:18
- The power to know d_labes 2008-05-09 09:35
- The power to know Nirali 2008-05-09 11:00
- The power to know d_labes 2008-05-16 08:54
- The power to know Nirali 2008-05-09 11:00
- The power to know d_labes 2008-05-09 09:35
- The power to know kevan 2009-05-25 15:46
- Bogus statement for 2,2,2-BE ElMaestro 2009-05-25 22:25
- Bogus? What? d_labes 2009-05-27 08:57
- Linear model on steroids ElMaestro 2009-05-28 19:12
- Bogus? What? d_labes 2009-05-27 08:57
- Fixed nowadays what? d_labes 2009-05-27 09:03
- Bogus statement for 2,2,2-BE ElMaestro 2009-05-25 22:25
- The power to know Helmut 2008-05-08 17:18
- Kinetica 5.0 bug Helmut 2008-12-31 16:42
- Error tests between/within Helmut 2008-05-06 21:33
- Error factor Ohlbe 2008-05-06 19:25
- Sample data & results Eva 2008-05-06 17:42