[R] testing differences between slope differences with lme

Dede Greenstein dede01 at codon.nih.gov
Tue Apr 10 01:18:54 CEST 2007


i have a mixed effect model which gives slope and intercept terms for 6 
groups  (diagnosis (3 levels) by risk group(2 levels)). the fixed part of 
the model is  --

brain volume ~ Diagnosis + Risk Group + (Risk Group * age : Diagnosis) - 1

thus allowing risk group age/slope terms to vary within diagnosis and 
omitting a nonsignificant diagnosis by risk group intercept (age was 
centered) interaction.

i am interested in whether differences in risk groups' developmental 
trajectories are different for different diagnoses.

the last three (of 10) fixed effect estimates are estimates for the 
age/slope differences between risk groups for the 3 diagnostic groups.

my question is this -- how do i test if the 3 differences are statistically 
different from each other?

i can do this in SAS but the answer in R is eluding me.

thanks in advance,

More information about the R-help mailing list