[R] Coefficients and standard errors in lme

Sebastian Luque sluque at mun.ca
Tue Mar 23 18:36:40 CET 2004


Hello,

I have been searching for ways to obtain these for combinations of fixed 
factors and levels other than the 'baseline' group (contrasts coded all 
0's) from a mixed-effects model in lme. I've modelled the continuous 
variable y as a function of a continuous covariate x, and fixed factors 
A, B, and C. The fixed factors have two levels each and I'd like to know 
whether the relationship between y and x varies between levels of the 
factors, and whether there are any interactions between these factors. 
I've therefore setup the model as this:

lme.fit <- lme(y ~ x*A*B*C, data=df, random=~x | subjectID)

The contrasts are default ("contr.treatment" and "contr.poly"). As 
usual, the summary provides the coefficients for the 'baseline' group. 
The rest of coefficients correspond to *differences* and their standard 
error with respect to this group. One can calculate the coefficients for 
any combination of factor levels by adding the appropriate coefficients 
in the results. However, I don't understand how to obtain the standard 
errors for these from the summary report. Can someone please let me know 
how to obtain these?

Cheers,
Sebastian




More information about the R-help mailing list