Hi,
I have a data set containing: group a, group b and group c. All subjects
perform task 1, 2, 3 and rest.
my matrix is set up:
task 1/ev1: 1 0 0
task 2/ev2: 0 1 0
task 3/ev3: 0 0 1
which is fine, but...
1>3? 1 0 -1
3>1? -1 0 1
and so on...
Q's: Doesn't the matrix above model linear increases? How do I set it up to
answer the question "where does for example 3 activate more than 1"? Can I
only model 2 ev's at a time, and then have to perform a higher-level
analysis to check for "1>3"?
Kind regards,
Benny
Karolinska Institute
Clinical CNS
|