This is a question about lme()
syntax. My response variable is 'response'. My fixed variable is 'year'. I have 2 random variables: 'student' which is nested within 'school'.
I want to include a year*school
interaction, but I do NOT want to include a year*student
one.
This is the syntax I have so far, but this seems to include 2 random interactions where I only want 1.
lme1 = lme(response ~ 1 + year, random = ~ year | school/student,
method = "REML", data = data)
This would be much easier to do in lmer
from the lme4
package. Assuming each student has a unique identifier, it would look like this (untested).
lmer(response ~ year + (year | school) + (1 | student), data = data)