[R] calculate the elasticities by linear.hypothesi commander

[John Fox]

Dear Jinghua Xie,

This is, I assume, the linear.hypothesis function in the car package, which
is deprecated in favour of linearHypothesis.

Please see below:


> -----Original Message-----
> From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org]
On
> Behalf Of Jinghua Xie
> Sent: August-27-10 8:26 AM
> To: r-help at r-project.org
> Subject: [R] calculate the elasticities by linear.hypothesi commander
> 
> Dear all
> 
> 
> 
> If I run the model and get the  estimated parameter a11.  Then I want to
use
> the estimated parameter to calculate the elasticities by using the formula
> e11=a11/mw1-1.  What I have done is using the command of linear.
> Hypothesis.
> 
> 
> 
> > formulas1=dWfresh~dlnPfresh+dlnPfrozen+dlnPsmoke+dlnQP+cosL1+sinL1+cos
> > L2
> 
> > formulas2=dWsmoke~dlnPfresh+dlnPfrozen+dlnPsmoke+dlnQP+cosL1+sinL1+cos
> > L2
> 
> > formulas12=list(formulas1,formulas2)
> 
> > > salmon1=systemfit(formulas12, data=Fsalmon2, method="SUR")
> 
> > summary (salmon1)
> 
> 
> 
> >MWfresh<-mean (Wfresh2, data=Fsalmon)
> 
> >MWfrozen<-mean (Wfrozen2, data=Fsalmon)
> 
> >MWsmoke<-mean (Wsmoke2, data=Fsalmon)
> 
> 
> 
> >restricte11=c("eq1_dlnPfresh/MWfresh-1=0")
> 
> 
> 
> >linear.hypothesis(salmonsdlnf1r, restricte11, test="Chisq")
> 
> 
> 
> It does not work. However, if I replace MWfresh with the pure numerical
> number
> 
> restricte11=c("eq1_dlnPfresh/0.42-1=0")
> 
> 
> 
> It works.

linearHypothesis interprets names such as eq1_dlnPfresh and MWfresh as
coefficient names, so you have to use numeric constants; I don't think
"eq1_dlnPfresh/0.42-1=0" will work, however, because division isn't allowed;
you would have to rephrase as "2.38*eq1_dlnPfresh - 1". The deltaMethod
function, also in the car package, is more flexible.

I hope this helps,
 John

--------------------------------
John Fox
Senator William McMaster 
  Professor of Social Statistics
Department of Sociology
McMaster University
Hamilton, Ontario, Canada
web: socserv.mcmaster.ca/jfox
> 
> 
> 
> Can anyone help me to work it out.?
> 
> 
> 
> Thanks
> 
> 
> 
> Jinghua Xie
> 
> University of Tromsx, Norway
> 
> 
> 
> 
> 
> 
> 	[[alternative HTML version deleted]]