

Hello,
We use drc to fit doseresponse curves, recently we discovered that
there are quite different standard error values returned for the same
dataset depending on the drcversion / Rversion that was used (not
clear which factor is important)
On R 2.9.0 using drc_1.63 we get an IC50 of 1.27447 and a standard
error on the IC50 of 0.43540
Whereas on R 2.7.0 using drc_1.42 the IC50 is 1.2039e+00 and the
standard error is 3.7752e03
Normally I would use the most recent version (both R and drc library)
but it seems to me that a standard error of 0.4 on a mean of 1.2 is too
big, so I trust the values we get with the older versions more
Has anyone suggestions on
 how to solve these discrepancies, if possible
 how to calculate which one of the 2 solutions is the correct one?
Thanks a lot,
Hans Vermeiren
Demo (on a windows machine, while the issue was actually discovered on
our ubuntu linux server):
1)
sessionInfo()
R version 2.7.0 (20080422)
i386pcmingw32
locale:
LC_COLLATE=Dutch_Belgium.1252;LC_CTYPE=Dutch_Belgium.1252;LC_MONETARY=Du
tch_Belgium.1252;LC_NUMERIC=C;LC_TIME=Dutch_Belgium.1252
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] drc_1.42 plotrix_2.42 nlme_3.189 MASS_7.241
lattice_0.176
[6] alr3_1.1.7
loaded via a namespace (and not attached):
[1] grid_2.7.0
d<data.frame(dose=c(2.00e05, 4.00e06, 8.00e07, 1.60e07, 3.20e08,
6.40e09, 1.28e09, 2.56e10, 5.10e11, 1.00e11, 2.00e05, 4.00e06,
8.00e07, 1.60e07, 3.20e08, 6.40e09, 1.28e09, 2.56e10, 5.10e11,
1.00e11),
response=c(97.202,81.670,47.292,16.924, 16.832, 6.832, 11.118,
1.319, 5.495, 3.352, 102.464, 83.114, 50.631, 22.792, 18.348,
19.066, 27.794, 14.682, 11.992, 12.868))
m< drm(response ~ (log10(dose*1e6)), data = d, fct = l4(fixed =
c(NA,NA,NA,NA), names = c("hs", "bottom", "top", "ec50")), logDose = 10,
control = drmc(useD = T))
summary(m)
results in:
Model fitted: Loglogistic (ED50 as parameter) (4 parms)
Parameter estimates:
Estimate Std. Error tvalue pvalue
hs:(Intercept) 9.8065e01 2.5821e03 3.7979e+02 2.248e33
bottom:(Intercept) 1.0955e+01 2.2546e02 4.8591e+02 4.364e35
top:(Intercept) 1.0502e+02 9.0935e02 1.1549e+03 4.210e41
ec50:(Intercept) 1.2039e+00 3.7752e03 3.1890e+02 3.681e32
Residual standard error: 7.026655 (16 degrees of freedom)
========================================================================
===========================================
2)
sessionInfo()
R version 2.9.0 (20090417)
i386pcmingw32
locale:
LC_COLLATE=Dutch_Belgium.1252;LC_CTYPE=Dutch_Belgium.1252;LC_MONETARY=Du
tch_Belgium.1252;LC_NUMERIC=C;LC_TIME=Dutch_Belgium.1252
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] drc_1.63 plotrix_2.55 nlme_3.190 MASS_7.246
magic_1.44 abind_1.10 lattice_0.1722 alr3_1.1.7
loaded via a namespace (and not attached):
[1] grid_2.9.0 tools_2.9.0
d<data.frame(dose=c(2.00e05, 4.00e06, 8.00e07, 1.60e07, 3.20e08,
6.40e09, 1.28e09, 2.56e10, 5.10e11, 1.00e11, 2.00e05, 4.00e06,
8.00e07, 1.60e07, 3.20e08, 6.40e09, 1.28e09, 2.56e10, 5.10e11,
1.00e11),
response=c(97.202,81.670,47.292,16.924, 16.832, 6.832, 11.118,
1.319, 5.495, 3.352, 102.464, 83.114, 50.631, 22.792, 18.348,
19.066, 27.794, 14.682, 11.992, 12.868))
m< drm(response ~ (log10(dose*1e6)), data = d, fct = l4(fixed =
c(NA,NA,NA,NA), names = c("hs", "bottom", "top", "ec50")), logDose = 10,
control = drmc(useD = T))
summary(m)
gives:
Model fitted: Loglogistic (ED50 as parameter) (4 parms)
Parameter estimates:
Estimate Std. Error tvalue pvalue
hs:(Intercept) 0.95266 0.25778 3.69564 0.0020
bottom:(Intercept) 10.97437 2.24421 4.89009 0.0002
top:(Intercept) 106.38373 9.98378 10.65565 1.127e08
ec50:(Intercept) 1.27447 0.43540 2.92712 0.0099
Residual standard error:
7.020175 (16 degrees of freedom)

This email and its attachment(s) (if any) may contain confidential and/or proprietary information and is intended for its addressee(s) only. Any unauthorized use of the information contained herein (including, but not limited to, alteration, reproduction, communication, distribution or any other form of dissemination) is strictly prohibited. If you are not the intended addressee, please notify the orginator promptly and delete this email and its attachment(s) (if any) subsequently.
Galapagos nor any of its affiliates shall be liable for direct, special, indirect or consequential damages arising from alteration of the contents of this message (by a third party) or as a result of a virus being passed on.
______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/rhelpPLEASE do read the posting guide http://www.Rproject.org/postingguide.htmland provide commented, minimal, selfcontained, reproducible code.


On May 20, 2009, at 11:20 AM, Hans Vermeiren wrote:
> Hello,
>
> We use drc to fit doseresponse curves, recently we discovered that
> there are quite different standard error values returned for the same
> dataset depending on the drcversion / Rversion that was used (not
> clear which factor is important)
> On R 2.9.0 using drc_1.63 we get an IC50 of 1.27447 and a standard
> error on the IC50 of 0.43540
> Whereas on R 2.7.0 using drc_1.42 the IC50 is 1.2039e+00 and the
> standard error is 3.7752e03
> Normally I would use the most recent version (both R and drc library)
> but it seems to me that a standard error of 0.4 on a mean of 1.2 is
> too
> big, so I trust the values we get with the older versions more
> Has anyone suggestions on
>  how to solve these discrepancies, if possible
>  how to calculate which one of the 2 solutions is the correct one?
>
> Thanks a lot,
> Hans Vermeiren
>
> Demo (on a windows machine, while the issue was actually discovered on
> our ubuntu linux server):
> 1)
> sessionInfo()
> R version 2.7.0 (20080422)
> i386pcmingw32
>
> locale:
> LC_COLLATE=Dutch_Belgium.1252;LC_CTYPE=Dutch_Belgium.
> 1252;LC_MONETARY=Du
> tch_Belgium.1252;LC_NUMERIC=C;LC_TIME=Dutch_Belgium.1252
>
> attached base packages:
> [1] stats graphics grDevices utils datasets methods base
>
>
> other attached packages:
> [1] drc_1.42 plotrix_2.42 nlme_3.189 MASS_7.241
> lattice_0.176
> [6] alr3_1.1.7
>
> loaded via a namespace (and not attached):
> [1] grid_2.7.0
>
> d<data.frame(dose=c(2.00e05, 4.00e06, 8.00e07, 1.60e07, 3.20e08,
> 6.40e09, 1.28e09, 2.56e10, 5.10e11, 1.00e11, 2.00e05, 4.00e06,
> 8.00e07, 1.60e07, 3.20e08, 6.40e09, 1.28e09, 2.56e10, 5.10e11,
> 1.00e11),
> response=c(97.202,81.670,47.292,16.924, 16.832, 6.832, 11.118,
> 1.319, 5.495, 3.352, 102.464, 83.114, 50.631, 22.792, 18.348,
> 19.066, 27.794, 14.682, 11.992, 12.868))
>
> m< drm(response ~ (log10(dose*1e6)), data = d, fct = l4(fixed =
> c(NA,NA,NA,NA), names = c("hs", "bottom", "top", "ec50")), logDose =
> 10,
> control = drmc(useD = T))
>
> summary(m)
> results in:
> Model fitted: Loglogistic (ED50 as parameter) (4 parms)
>
> Parameter estimates:
>
> Estimate Std. Error tvalue pvalue
> hs:(Intercept) 9.8065e01 2.5821e03 3.7979e+02 2.248e33
> bottom:(Intercept) 1.0955e+01 2.2546e02 4.8591e+02 4.364e35
> top:(Intercept) 1.0502e+02 9.0935e02 1.1549e+03 4.210e41
> ec50:(Intercept) 1.2039e+00 3.7752e03 3.1890e+02 3.681e32
>
> Residual standard error: 7.026655 (16 degrees of freedom)
> =
> =
> ======================================================================
> ===========================================
> 2)
> sessionInfo()
> R version 2.9.0 (20090417)
> i386pcmingw32
>
> locale:
> LC_COLLATE=Dutch_Belgium.1252;LC_CTYPE=Dutch_Belgium.
> 1252;LC_MONETARY=Du
> tch_Belgium.1252;LC_NUMERIC=C;LC_TIME=Dutch_Belgium.1252
>
> attached base packages:
> [1] stats graphics grDevices utils datasets methods base
>
>
> other attached packages:
> [1] drc_1.63 plotrix_2.55 nlme_3.190 MASS_7.246
> magic_1.44 abind_1.10 lattice_0.1722 alr3_1.1.7
>
> loaded via a namespace (and not attached):
> [1] grid_2.9.0 tools_2.9.0
>
> d<data.frame(dose=c(2.00e05, 4.00e06, 8.00e07, 1.60e07, 3.20e08,
> 6.40e09, 1.28e09, 2.56e10, 5.10e11, 1.00e11, 2.00e05, 4.00e06,
> 8.00e07, 1.60e07, 3.20e08, 6.40e09, 1.28e09, 2.56e10, 5.10e11,
> 1.00e11),
> response=c(97.202,81.670,47.292,16.924, 16.832, 6.832, 11.118,
> 1.319, 5.495, 3.352, 102.464, 83.114, 50.631, 22.792, 18.348,
> 19.066, 27.794, 14.682, 11.992, 12.868))
>
> m< drm(response ~ (log10(dose*1e6)), data = d, fct = l4(fixed =
> c(NA,NA,NA,NA), names = c("hs", "bottom", "top", "ec50")), logDose =
> 10,
> control = drmc(useD = T))
>
> summary(m)
>
> gives:
> Model fitted: Loglogistic (ED50 as parameter) (4 parms)
>
> Parameter estimates:
>
> Estimate Std. Error tvalue pvalue
> hs:(Intercept) 0.95266 0.25778 3.69564 0.0020
> bottom:(Intercept) 10.97437 2.24421 4.89009 0.0002
> top:(Intercept) 106.38373 9.98378 10.65565 1.127e08
> ec50:(Intercept) 1.27447 0.43540 2.92712 0.0099
>
> Residual standard error:
>
> 7.020175 (16 degrees of freedom)
Hans,
You have three important factors changing here. The version of R, the
version of drc and the versions of any relevant drc dependencies
(alr3, lattice, magic, MASS, nlme, plotrix).
I would first try to install the newer version of drc on the older R
system (all else staying the same) and see what you get. Don't run
update.packages() here, lest you change other things. Just install the
newer version of drc.
If you get the same results as the older version, then it might lead
you to something in R or one of the package dependencies changing.
If you get a different result, then it would lead to something in drc
changing.
You can also install the old version of drc on your more recent R
system to see what you get, which might help to confirm behavior.
The old source version of drc would be available from:
http://cran.rproject.org/src/contrib/Archive/drc/I also found a Windows binary of the old package here:
http://cran.rproject.org/bin/windows/contrib/2.6/drc_1.42.zipI have also copied Christian Ritz, the drc package author/maintainer,
so that he may be able to assist you further with the problem.
HTH,
Marc Schwartz
______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/rhelpPLEASE do read the posting guide http://www.Rproject.org/postingguide.htmland provide commented, minimal, selfcontained, reproducible code.


Hello
Thanks a lot Marc, for the suggestion to explore the issue a bit more
systematically
So I did and the conclusion is that with the old drc 1.42, I get a
SE=0.003, with the new drc 1.52, I get a SE=0.4
irrespective of the R version or the version of the packages drc depends
on
I hope somebody can help me further, I still have the feeling that a SE
of 0.003 on an IC50 of 1.2, for a reasonable good fit is more realistic
than a SE of 0.4 on an IC50 of 1.2
Regards,
Hans
Here are the results of the following test script:
d<data.frame(dose=c(2.00e05,4.00e06,8.00e07,1.60e07,3.20e08,6.40e
09,1.28e09,2.56e10,5.10e11,1.00e11,2.00e05,4.00e06,8.00e07,1.60e
07,3.20e08,6.40e09,1.28e09,2.56e10,5.10e11,1.00e11),response=c(97.
202,81.670,47.292,16.924, 16.832, 6.832, 11.118, 1.319, 5.495,
3.352, 102.464, 83.114, 50.631, 22.792, 18.348, 19.066, 27.794,
14.682, 11.992, 12.868))
m< drm(response ~ (log10(dose*1e6)), data = d, fct = l4(fixed =
c(NA,NA,NA,NA), names = c("hs", "bottom", "top", "ec50")), logDose = 10,
control = drmc(useD = T))
summary(m)
RESULTS:
sessionInfo()
R version 2.7.0 (20080422)
i386pcmingw32
locale:
LC_COLLATE=Dutch_Belgium.1252;LC_CTYPE=Dutch_Belgium.1252;LC_MONETARY=Du
tch_Belgium.1252;LC_NUMERIC=C;LC_TIME=Dutch_Belgium.1252
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] drc_1.52 plotrix_2.42 nlme_3.189 MASS_7.241
lattice_0.176
[6] alr3_1.1.7
loaded via a namespace (and not attached):
[1] grid_2.7.0 tools_2.7.0
RESULT: ec50:(Intercept)=1.27447 SE=0.43541
CONCLUSION: R 2.7.0 with recent drc 1.52 (older dependencies) gives
SE=0.43541
========================================================================
=======
R version 2.9.0 (20090417)
i386pcmingw32
locale:
LC_COLLATE=Dutch_Belgium.1252;LC_CTYPE=Dutch_Belgium.1252;LC_MONETARY=Du
tch_Belgium.1252;LC_NUMERIC=C;LC_TIME=Dutch_Belgium.1252
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] drc_1.42 plotrix_2.55 nlme_3.190 MASS_7.246
[5] lattice_0.1722 alr3_1.1.7
loaded via a namespace (and not attached):
[1] grid_2.9.0 tools_2.9.0
RESULT: ec50:(Intercept) SE=1.2039e+00
CONCLUSION: R 2.9.0 with old drc 1.42 (newer dependencies) gives
SE=0.003
Hans,
You have three important factors changing here. The version of R, the
version of drc and the versions of any relevant drc dependencies
(alr3, lattice, magic, MASS, nlme, plotrix).
I would first try to install the newer version of drc on the older R
system (all else staying the same) and see what you get. Don't run
update.packages() here, lest you change other things. Just install the
newer version of drc.
If you get the same results as the older version, then it might lead
you to something in R or one of the package dependencies changing.
If you get a different result, then it would lead to something in drc
changing.
You can also install the old version of drc on your more recent R
system to see what you get, which might help to confirm behavior.
The old source version of drc would be available from:
http://cran.rproject.org/src/contrib/Archive/drc/I also found a Windows binary of the old package here:
http://cran.rproject.org/bin/windows/contrib/2.6/drc_1.42.zipI have also copied Christian Ritz, the drc package author/maintainer,
so that he may be able to assist you further with the problem.
HTH,
Marc Schwartz

This email and its attachment(s) (if any) may contain confidential and/or proprietary information and is intended for its addressee(s) only. Any unauthorized use of the information contained herein (including, but not limited to, alteration, reproduction, communication, distribution or any other form of dissemination) is strictly prohibited. If you are not the intended addressee, please notify the orginator promptly and delete this email and its attachment(s) (if any) subsequently.
Galapagos nor any of its affiliates shall be liable for direct, special, indirect or consequential damages arising from alteration of the contents of this message (by a third party) or as a result of a virus being passed on.
______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/rhelpPLEASE do read the posting guide http://www.Rproject.org/postingguide.htmland provide commented, minimal, selfcontained, reproducible code.


Hi Hans,
I hope I can resolve your problems below (Marc, thank you very much for cc'ing me on your
initial response!).
Have a look at the following R lines:
## Fitting the model using drm() (from the latest version)
m1< drm(response ~ dose, data = d, fct = LL.4())
summary(m1)
plot(m1)
## Checking the fit by using nls()
## (we have very good guesses for the parameter estimates)
m2 < nls(response ~ c + (d  c)/(1 + (dose/e)^b), data=d, start=list(b=0.95, c=10,
d=106, e=1.2745e06))
summary(m2)
The standard errors agree quite well. The minor discrepancies between to two fits are
attributable to different numerical approximations of the variancecovariance matrix being
used in drm() and nls().
So I would use the latest version of 'drc', especially for datasets with really small
doses. One recent change to drm() was to incorporate several layers of scaling prior to
estimation (as well as subsequent back scaling after estimation):
1) scaling of parameters with the same scale as the x axis
2) scaling of parameters with the same scale as the y axis
3) scaling of parameters in optim()
The effect of scaling is to temporarily "convert" the dataset (and the model) to scales
that are more convenient for the estimation procedure. Any feedback on this would be much
appreciated.
Therefore it should also not be necessary to manually do any scaling prior to using drm()
(like what you did). Compare, for instance, your specification of drm() to mine above.
Is this explanation useful?!
Christian
______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/rhelpPLEASE do read the posting guide http://www.Rproject.org/postingguide.htmland provide commented, minimal, selfcontained, reproducible code.


Yes, thanks that's very useful
Apart from checking the fit with nls() as you suggested, I've also used Prism, which gave the following results
Equation 1
Bestfit values
BOTTOM 10.96
TOP 106.4
LOGEC50 5.897
HILLSLOPE 0.9501
EC50 1.2670e006
Std. Error
BOTTOM 2.196
TOP 9.337
LOGEC50 0.1439
HILLSLOPE 0.2270
95% Confidence Intervals
BOTTOM 6.301 to 15.61
TOP 86.62 to 126.2
LOGEC50 6.202 to 5.592
HILLSLOPE 0.4689 to 1.431
EC50 6.2750e007 to 2.5560e006
Goodness of Fit
Degrees of Freedom 16
R² 0.9622
Absolute Sum of Squares 787.5
Sy.x 7.015
Data
Number of X values 20
Number of Y replicates 1
Total number of values 20
Number of missing values 0
In other words: also in line with the drc 1.63 and nls() results
As for the scaling: yes this is useful because I can't predict whether concentrations are in molar, micromolar,..., right now I indeed scaled dosevalues "manually", it's better/ more robust when the drmfunction takes care of that
I suppose this also means I don't have to do the log transformation anymore?
Thanks (both of you) for your swift feedback
Hans
Original Message
From: Christian Ritz [mailto: [hidden email]]
Sent: vrijdag 22 mei 2009 11:30
To: Hans Vermeiren
Cc: [hidden email]; [hidden email]
Subject: Re: [R] drc results differ for different versions
Hi Hans,
I hope I can resolve your problems below (Marc, thank you very much for cc'ing me on your
initial response!).
Have a look at the following R lines:
## Fitting the model using drm() (from the latest version)
m1< drm(response ~ dose, data = d, fct = LL.4())
summary(m1)
plot(m1)
## Checking the fit by using nls()
## (we have very good guesses for the parameter estimates)
m2 < nls(response ~ c + (d  c)/(1 + (dose/e)^b), data=d, start=list(b=0.95, c=10,
d=106, e=1.2745e06))
summary(m2)
The standard errors agree quite well. The minor discrepancies between to two fits are
attributable to different numerical approximations of the variancecovariance matrix being
used in drm() and nls().
So I would use the latest version of 'drc', especially for datasets with really small
doses. One recent change to drm() was to incorporate several layers of scaling prior to
estimation (as well as subsequent back scaling after estimation):
1) scaling of parameters with the same scale as the x axis
2) scaling of parameters with the same scale as the y axis
3) scaling of parameters in optim()
The effect of scaling is to temporarily "convert" the dataset (and the model) to scales
that are more convenient for the estimation procedure. Any feedback on this would be much
appreciated.
Therefore it should also not be necessary to manually do any scaling prior to using drm()
(like what you did). Compare, for instance, your specification of drm() to mine above.
Is this explanation useful?!
Christian

This email and its attachment(s) (if any) may contain confidential and/or proprietary information and is intended for its addressee(s) only. Any unauthorized use of the information contained herein (including, but not limited to, alteration, reproduction, communication, distribution or any other form of dissemination) is strictly prohibited. If you are not the intended addressee, please notify the orginator promptly and delete this email and its attachment(s) (if any) subsequently.
Galapagos nor any of its affiliates shall be liable for direct, special, indirect or consequential damages arising from alteration of the contents of this message (by a third party) or as a result of a virus being passed on.
______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/rhelpPLEASE do read the posting guide http://www.Rproject.org/postingguide.htmland provide commented, minimal, selfcontained, reproducible code.


I've been having trouble making sense of the drc results for my binomial response toxicity data. Firstly, the standard errors are far too large for how well the data fit the loglogistic model, particularly compared to other methods of LC50 estimation (e.g. probit and trimmed spearmankarber). Secondly, the LC50 estimates produced do not even fit with the DRC curve (see example below). I've tried all sort of loglogistic functions with varying fixed parameters (i.e. upper and lower limits for mortality data) and still fail to produce reliable estimates. Furthermore, my data response curves exhibit asymmetry so I feel the need to use the 5 parm function. There seems to be some inconsistency here which is causing me a serious headache.
#data 1#
n<c(20,20,20,20,20,20,20,20,20,20,20,20,20,20,20,20,20,20,20,20,20)
x<c(256,256,256,847,847,847,2208,2208,2208,3565,3565,3565,5756,5756,5756,7313,7313,7313,15000,15000,15000)
r<c(0,0,0,0,2,0,1,3,4,7,6,9,17,13,18,18,17,20,20,20,20)
p<r/n
d<data.frame(x=x,n=n,r=r,p=p)
fct < LL2.5(fixed = c(NA,0,1,NA,NA),names = c("Hillslope", "Lower Limit", "Upper Limit", "LC50", "Asymmetry"))
drc < drm(p~log(x), weight=n, data=d, fct=fct, type="binomial", logDose=exp(1))
summary(drc)
plot(drc)
est1<8.67
se1<0.14
points(est1,0.5)
arrows(est1,0.5,est1se1,angle=90,code=2,length=0.025)
arrows(est1,0.5,est1+se1,angle=90,code=2,length=0.025)
#data 2#
n<c(20,20,20,20,20,20,20,20,20,20,20,20,20,20,20,20,20,20)
x<c(798,798,798,6770,6770,6770,12130,12130,12130,21736,21736,21736,38948,38948,38948,69791,69791,69791)
r<c(0,0,0,4,3,3,5,8,7,12,12,7,17,15,17,20,20,19)
p<r/n
d2<data.frame(x=x,n=n,r=r)
d2
fct < LL2.5(fixed = c(NA,0,1,NA,NA),names = c("Hillslope", "Lower Limit", "Upper Limit", "LC50", "Asymmetry"))
drc2 < drm(p~log(x), weight=n, data=d2, fct=fct, type="binomial", logDose=exp(1))
summary(drc2)
plot(drc2)
est2<10.56
se2<0.271
points(est2,0.5)
arrows(est2,0.5,est2se2,angle=90,code=2,length=0.025)
arrows(est2,0.5,est2+se2,angle=90,code=2,length=0.025)
The idea is to use the drc package for these types of data to produce isoboles. Has anyone else encountered this issue before and can help me? OR I noticed Dr. Ritz has responded to this thread, do you have an suggestions? As far as I know, I have the latest version of every package for R.
Sincerely,
Patrick


P.S.
Here are the results I get for;
#data 1#
Model fitted: Generalised loglogistic (log(ED50) as parameter) (3 parms)
Parameter estimates:
Estimate Std. Error tvalue pvalue
Hillslope:(Intercept) 6.05453 2.18455 2.77152 0.0056
LC50:(Intercept) 8.67707 0.14016 61.90663 0.0000
Asymmetry:(Intercept) 0.32382 0.16608 1.94978 0.0512
#data 2#
Model fitted: Generalised loglogistic (log(ED50) as parameter) (3 parms)
Parameter estimates:
Estimate Std. Error tvalue pvalue
Hillslope:(Intercept) 3.98218 1.91035 2.08453 0.0371
LC50:(Intercept) 10.55736 0.27120 38.92901 0.0000
Asymmetry:(Intercept) 0.26840 0.18244 1.47120 0.1412
Pat

