R help - Jul 2004 - confidence intervals for linear combinations when using lme

Hi 

I really hope someone can help me.

I have just started to work with S-plus, and have not yet understood how it
really works. I am now trying to fit a mixed effects model with lme. My goal
is to compare four different groups, at several different time points, and I
therefore would like to create confidence intervals for linear combinations
of my estimated parameters (as I usually do with "contrast" or
"estimate" or
"lsmeans" in SAS). I have now found out that the vcov-function in the
lme4-package could be of great help for me, but I do not know if I can use
lme4 when working with S-PLUS 6.1, or is it only for R. 

When I write 
>library(MASS)
I find lme3, where there is a vcov-function, but if I understand it
correctly it works with lm and not with lme...

If I can use lme4 with s-plus 6.1, what shall I do in order to install it on
my computer?

I am grateful for all help I can get?

best regards
Anna Persson

On Friday 23 July 2004 05:36, Anna.H.Persson at astrazeneca.com
wrote:
> Hi
>
> I really hope someone can help me.
>
> I have just started to work with S-plus, and have not yet understood
> how it really works. I am now trying to fit a mixed effects model
> with lme. My goal is to compare four different groups, at several
> different time points, and I therefore would like to create
> confidence intervals for linear combinations of my estimated
> parameters (as I usually do with "contrast" or
"estimate" or
> "lsmeans" in SAS). I have now found out that the vcov-function in
the
> lme4-package could be of great help for me, but I do not know if I
> can use lme4 when working with S-PLUS 6.1, or is it only for R.
As far as I know, there are no plans (by the developers) to port lme4 to 
work with S-PLUS.

Deepayan

Anna.H.Persson at astrazeneca.com wrote:
> I have just started to work with S-plus, and have not yet understood how it
> really works. I am now trying to fit a mixed effects model with lme. My
goal
> is to compare four different groups, at several different time points, and
I
> therefore would like to create confidence intervals for linear combinations
> of my estimated parameters (as I usually do with "contrast" or
"estimate" or
> "lsmeans" in SAS). I have now found out that the vcov-function in
the
> lme4-package could be of great help for me, but I do not know if I can use
> lme4 when working with S-PLUS 6.1, or is it only for R. 
The lme4 package is only available for R at this time.