Workarounds for v6.3+ [Software]
Hi Mittyri and Antigoni,
Confirmed. I checked the 64bit-versions of PHX/WNL 6.3, 6.4, 7.0, and 8.0 (Linear Mixed Effects and WNL Linear Model 502, 1–7 degrees of freedom). The critical t-value is not reported in Model 502. But we can calculate it from the confidence limit(s), the estimate, and its standard error as
Interesting Scary that the reported critical t-value in Linear Mixed Effects is also wrong! I didn’t expect that… Results are identical (full numeric precision) in all versions of PHX/WNL. The custom function
Given that, use only LME (map time as Regressor, concentration as Dependent, drag time from Regressors/Covariates to Model Specification). After execution I suggest these workarounds:
In ≥6.4: Results > Final Fixed Parameters > Send To > Data > Data Wizard > Add > Transformation >
Transformation Type: Custom
New Column Name:
Formula:
Add another one named In BE we don’t need the CI of fixed effects and the degrees of freedom are generally high. However, I’m dumbfounded by such defects in commercial software. Even for high degrees of freedom
If you don’t trust in the open source software I used, below an excerpt of R.A. Fisher’s Statistical Methods for Research Workers (1934), the Geigy tables (1980), WolframAlpha, and UsableStats:
❝ As I see the problem is in quantile function, looks like approximation is not correct for low DF
Confirmed. I checked the 64bit-versions of PHX/WNL 6.3, 6.4, 7.0, and 8.0 (Linear Mixed Effects and WNL Linear Model 502, 1–7 degrees of freedom). The critical t-value is not reported in Model 502. But we can calculate it from the confidence limit(s), the estimate, and its standard error as
(UnivarCI_Upper – Estimate) / StdError
or (Estimate – UnivarCI_Lower) / StdError
.tinv(0.025, df)
introduced in v6.4 is practically correct.df R/OO/Gnumeric LME (rep’d) %RE 502 (calc) %RE tinv(0.025, df) %RE
1 12.706204736175 11.300117888951 –11.0661 16.500679825590 +29.8632 12.706204736175 ±0.0000
2 4.302652729749 4.270551416408 –0.7461 4.327589042282 +0.5796 4.302652729750 +0.0000
3 3.182446305284 3.178617037716 –0.1203 3.183479996442 +0.0325 3.182434925786 –0.0004
4 2.776445105198 2.775582256369 –0.0311 2.776407538431 –0.0014 2.776445033314 –0.0000
5 2.570581835636 2.570308254272 –0.0106 2.570544577883 –0.0014 2.570581685935 –0.0000
6 2.446911851145 2.446804416220 –0.0044 2.446913617243 +0.0001 2.446911763977 –0.0000
7 2.364624251593 2.364575396335 –0.0021 2.364643234749 +0.0008 2.364624210734 –0.0000
Given that, use only LME (map time as Regressor, concentration as Dependent, drag time from Regressors/Covariates to Model Specification). After execution I suggest these workarounds:
- In v6.3 (might work in earlier versions as well; untested!).
- Data > New > Worksheet; rename to tcrit
Two columns:Denom_DF
andtinv
To get the critical t-values:
♦ R-code
for (df in 1:28) {
cat(sprintf("%i %17.14f", df, qt(0.025, df, lower.tail=FALSE)), "\n")
}
♦ Open Office Calc, Gnumeric, or current version of M$-Excel
column A: integer degrees of freedom starting with 1
name column A as df
column B:=TINV(0.05, df)
: drag down to the maximum DFs
♦ or enter them manually from UsableStats.
- Navigate to Linear Mixed Effects > Output Data > Final Fixed Parameters > Send To > Data > Merge Worksheets
Sort: Denom_DF
Included Column: Effect_Level, Estimate, StdError
Worksheet 2: tcrit
Sort: Denom_DF
Included Column: tinv
Execute.
- Output Data > Result > Send To > Data > Data Wizard
Add a filter: Action: Exclude, Find: NULL, Column: Effect_Level
Add a filter: Action: Exclude, Find: int, Column: Effect_Level
Add > Transformation >
Transformation Type: Custom
New Column Name:CL_lo
Formula:Estimate-tinv*StdError
Add > Transformation >
Transformation Type: Custom
New Column Name:CL_hi
Formula:Estimate+tinv*StdError
Transformation Type: Custom
New Column Name:
CL_lo
Formula:
Estimate-tinv(0.025, Denom_DF)*StdError
Add another one named
CL_hi
, Formula: Estimate+tinv(0.025, Denom_DF)*StdError
tinv(x, y)
is three orders of magnitude “better” than what is given by LME/BE. Why is it not used?df R/OO/Gnumeric LME = BE %RE tinv(0.025, 28) %RE
28 2.04840714179524 2.04840711367364 –1.37·10–6 2.04840714175209 –2.11·10–9
If you don’t trust in the open source software I used, below an excerpt of R.A. Fisher’s Statistical Methods for Research Workers (1934), the Geigy tables (1980), WolframAlpha, and UsableStats:
df t(p 0.025)
1 12.706 12.7062 12.706 12.70620474
2 4.303 4.3027 4.303 4.30265273
3 3.182 3.1824 3.182 3.18244912
4 2.776 2.7764 2.776 2.77644504
5 2.571 2.5706 2.571 2.57058169
6 2.447 2.4469 2.447 2.44691177
7 2.365 2.3646 2.365 2.36462421
28 2.048 2.0484 2.048 2.04840714
—
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!
![[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
Complete thread:
- assessment of steady state in WinNonLin pkpdpkpd 2007-04-04 22:52 [Software]
- steady state in WinNonlin (example) Helmut 2007-04-05 02:30
- steady state in WinNonlin (example) pkpdpkpd 2007-04-05 16:33
- steady state in WinNonlin (example) Helmut 2007-04-05 17:16
- steady state in WinNonlin (example) pkpdpkpd 2007-04-05 17:39
- steady state in WinNonlin… Helmut 2007-04-05 18:40
- steady state in WinNonlin… pkpdpkpd 2007-04-05 21:37
- Paired Design in WinNonlin Helmut 2007-04-06 00:23
- steady state in WinNonlin… Helmut 2007-04-05 18:40
- steady state in WinNonlin (example) pkpdpkpd 2007-04-05 17:39
- steady state in WinNonlin (example) Helmut 2007-04-05 17:16
- steady state in WinNonlin (example) apapanas 2018-01-03 14:44
- steady state in WinNonlin (example) Helmut 2018-01-04 17:11
- qt in WinNonlin mittyri 2018-01-04 23:13
- ‘Deep’ bug Helmut 2018-01-05 00:37
- Workarounds for v6.3+Helmut 2018-01-08 00:40
- Workarounds for v6.3+ apapanas 2018-01-10 12:47
- qt in WinNonlin mittyri 2018-01-04 23:13
- steady state in WinNonlin (example) Helmut 2018-01-04 17:11
- steady state in WinNonlin (example) pkpdpkpd 2007-04-05 16:33
- steady state in WinNonlin (example) Helmut 2007-04-05 02:30