R help - Aug 2006 - Setting contrasts for polr() to get same result of SAS

Hi all,

I am trying to do a ordered probit regression using polr(), replicating a
result from SAS.
>polr(y ~ x, dat, method='probit')
suppose the model is y ~ x, where y is a factor with 3 levels and x is a
factor with 5 levels,

To get coefficients, SAS by default use the last level as reference, R by
default use the first level (correct me if I was wrong),

The result I got is a mixture, using first and last for different variables.

I tried relevel, reorder, contrasts, but no success. I found what really
matters is
>options(contrasts = c("contr.treatment", "contr.poly"))
or
>options(contrasts = c("contr.SAS", "contr.poly"))
so I guess I can set "contrasts= a list of contrasts" for each
variables in
polr(), but I cannot successfuly set the contrasts, what is the syntax?

Is there a better way to do this?

Thank you very much,
Tian

	[[alternative HTML version deleted]]

T Mu <muster <at> gmail.com> writes:
> 
> Hi all,
> 
> I am trying to do a ordered probit regression using polr(), replicating a
> result from SAS.
> 
> >polr(y ~ x, dat, method='probit')
> 
> suppose the model is y ~ x, where y is a factor with 3 levels and x is a
> factor with 5 levels,
> 
> To get coefficients, SAS by default use the last level as reference, R by
> default use the first level (correct me if I was wrong),
Yes.
> I tried relevel, reorder, contrasts, but no success. I found what really
> matters is
I am sure those can help but you need to be carefull to "reorder"
levels
that the order is the same in SAS and R.
> >options(contrasts = c("contr.treatment",
"contr.poly"))
> 
> or
> 
> >options(contrasts = c("contr.SAS", "contr.poly"))
You can also set contrasts directly to factors via

contrasts(x) <- contr.SAS

where x is your factor. You can also set different contrasts to 
different factors.

Gregor