: Re: [R] lme - problems with model
To: CG Pettersson [EMAIL PROTECTED]
Cc: Douglas Bates [EMAIL PROTECTED], [EMAIL PROTECTED]
Message-ID: [EMAIL PROTECTED]
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
If you want to try to get the same answers as PROC MIXED, I
suggest you try
Thanks a lot for the answer!
Now, I only have the last one left - How do I get round it?
I knew about the missing cells in the design, but didn´t know how lme
would react on them.
In this case, I can remove the water:temp term, but how can I be sure
that this is the right thing to do?
Is the lm
If you want to try to get the same answers as PROC MIXED, I
suggest you try to figure out how SAS codes interactions and which ones
it retains. Then you can try code those manually and include them as
separate explanatory variables, e.g., I((water==2)(temp==110)). You
could work this
CG Pettersson [EMAIL PROTECTED] writes:
Thanks a lot for the answer!
Now, I only have the last one left - How do I get round it?
I knew about the missing cells in the design, but didn´t know how lme
would react on them.
In this case, I can remove the water:temp term, but how can I be sure
Doug's xtabs suggests to me that the following might be
estimable, with data.$Temp - as.numeric(as.character(data.$temp))
water*(Temp+I(Temp^2))
It looks like it should be estimable in lm, and depending on the
noise model, it should also be estimable in lme. ???
hope this
