Log in
RegisterSequence effect bear vs WinNonlin opposite results [Software]
Thank you all for your answers and excuse me for the delay of my reply
In request to ElMaestro and Helmut
here are my results from AUC0-t using WinNonlin 6.3.
I found out that WinNonlin estimate the F-value like this
Using Bear data and applying the previous formula I obtain the same result of WinNonlin
I think that is the root of the situation, and please correct me if I´m wrong.
Gracias!
In request to ElMaestro and Helmut
❝ if you paste the ANOVA tables from Winnonlin and Bear then it will be possible thorugh the MS to....determine...the source of this phenomenon....
here are my results from AUC0-t using WinNonlin 6.3.
- WinNonlin
- Response: Log10(AUCall)
Dependent | Units | Hypotesis | DF | SS | MS | F-value | P-value
-------------------------------------------------------------------------------------------
Log10(AUCall) | h*µg/mL | Sequence | 1 | 0.086471 | 0.086471| 4.348724 | 0.07053
Log10(AUCall) | h*µg/mL | Sequence*Volunteer | 8 | 0.019884 | 0.019884| 28.513921| 0.00004
Log10(AUCall) | h*µg/mL | Product | 1 | 0.014607 | 0.014067| 7.639860 | 0.02452
Log10(AUCall) | h*µg/mL | Period | 1 | 0.007480 | 0.007480| 3.912380 | 0.08331
Log10(AUCall) | h*µg/mL | Error | 8 | 0.015296 | 0.001912
- Bear
- Response:
Statistical analysis (ANOVA(lm))
-----------------------------------------------------------
Dependent Variable: log(AUC0t)
Analysis of Variance Table
Response: log(AUC0t)
Df Sum Sq Mean Sq F value Pr(>F)
prd 1 0.08335 0.083349 22.543 0.001449 **
drug 1 0.06330 0.063303 17.122 0.003263 **
subj(seq) 8 0.84340 0.105425 28.514 4.248e-05 ***
Residuals 8 0.02958 0.003697
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Analysis of Variance Table
Response: log(AUC0t)
DF Type III SS Mean Square F Value Pr > F
prd 1 0.08335 0.083349 22.543 0.0014490
drug 1 0.06330 0.063303 17.122 0.0032632
subj(seq) 8 0.84340 0.105425 28.514 0.0000425
-----------------------------------------------------------
Tests of Hypothesis using the Type I MS for
SUBJECT(SEQUENCE) as an error term
Df Sum Sq Mean Sq F value Pr(>F)
seq 1 0.4585 0.4585 8.093 0.0107 *
Residuals 18 1.0196 0.0566
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Tests of Hypothesis using the Type III MS for
SUBJECT(SEQUENCE) as an error term
Df Sum Sq Mean Sq F value Pr(>F)
seq 1 0.4585 0.4585 8.093 0.0107 *
Residuals 18 1.0196 0.0566
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1I found out that WinNonlin estimate the F-value like this
Sequence effect
F-value = MS sequence / MS sequence*volunteerUsing Bear data and applying the previous formula I obtain the same result of WinNonlin
F-value = 0.4585/0.105425 = 4.34.I think that is the root of the situation, and please correct me if I´m wrong.
Gracias!
Complete thread:
- Sequence effect bear vs WinNonlin opposite results Risherd 2012-08-27 20:39 [Software]
![[Mix]](https://static.bebac.at/img/mix.png "open in Mix view")
- Sequence effect bear vs WinNonlin opposite results yjlee168 2012-08-27 21:35
- Sequence effect bear vs WinNonlin opposite results ElMaestro 2012-08-27 22:37
- Sequence effect bear vs WinNonlin opposite results yjlee168 2012-08-29 02:53
- SAS Type I/III ↔ WinNonlin partial/sequential tests Helmut 2012-08-28 14:39
- Sequence effect bear vs WinNonlin opposite results ElMaestro 2012-08-28 19:13
- Sequence effect bear vs WinNonlin opposite resultsRisherd 2012-08-29 18:09
- Division by 0.0566? ElMaestro 2012-08-29 19:42
- Related to AUC Parameter… Tushar.g 2012-09-11 13:29
- Sequence effect bear vs WinNonlin opposite resultsRisherd 2012-08-29 18:09
Ing. Helmut Schütz