R help - Mar 2010 - testing parallelism of does-response curves using nls()

Hi, I am trying to use F test or Chi-square test to test if 2 5-parameter (A, B,
xmid, scal and H) logistic curves are parallel based on residual sum of squares.

What's usually done is to first fit the 2 curves using a constraint (or
global) model where all parameters are kept the same except for
"xmid"; then fit 2 independent curves using unconstraint models where
all 5 parameters are allowed for change. The extra residual sum of squares can
then be used to test parallelism using either Chi-square or F test.

Now, fitting 2 separate curves are straight forward using nls(), but how to fit
a global model to only allow "xmid" to be different among the 2
curves?

BTW, is there a selfStart function for 5-parameter logistic like SSfpl for
4-parameter logistic function?

Thanks very much,

John

Hi John,
you may have a look at the drc-package (or look at the maintainers web 
site www.bioassay.dk)
I think it does exactly what you want.
HTH
Eik

array chip schrieb:
> Hi, I am trying to use F test or Chi-square test to test if 2 5-parameter
(A, B, xmid, scal and H) logistic curves are parallel based on residual sum of
squares.
>
> What's usually done is to first fit the 2 curves using a constraint (or
global) model where all parameters are kept the same except for
"xmid"; then fit 2 independent curves using unconstraint models where
all 5 parameters are allowed for change. The extra residual sum of squares can
then be used to test parallelism using either Chi-square or F test.
>
> Now, fitting 2 separate curves are straight forward using nls(), but how to
fit a global model to only allow "xmid" to be different among the 2
curves?
>
> BTW, is there a selfStart function for 5-parameter logistic like SSfpl for
4-parameter logistic function?
>
> Thanks very much,
>
> John
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>