Hi Amanda,

Both are correct. The first will allow for age slopes to be different for each group, whereas the second forces the age slope to be the same for both groups. Even though you are not testing age effects, would you like to assume that age can affect differently each of the groups? If so, use the first model, otherwise use the second.

All the best,

Anderson


On Tue, 3 Sep 2019 at 14:25, Amanda Rodrigue <[log in to unmask]> wrote:
Hello,
I would like to test the interaction between diagnosis and trauma (continuous) on FA but I want to do so while controlling for the effects of age. Could someone tell me which of the following design matrices are correct? I am also demeaning variables across the whole sample before splitting them into group-wise columns.

option 1
EV1(grp1) EV2(grp 2) EV3(trauma grp1) EV4(trauma grp2) EV5(age grp1) EV6(age grp2)
1  0  -0.02   0     12.7    0
1  0  -0.43   0     15.7    0
1  0   0.05   0      8.7     0
0  1    0    -0.36    0   -13.2
0  1    0     0.62    0    -3.2
0  1    0     0.35    0    -3.2

contrast
0 0 1 -1 0 0
0 0 -1 1 0 0

OR

option 2
EV1(grp1) EV2(grp 2) EV3(trauma grp1) EV4(trauma grp2) EV5(age whole sample)
1  0  -0.02  0       12.7   
1  0  -0.43  0       15.7   
1  0  0.05   0        8.7     
0  1    0    -0.36  -13.2
0  1    0     0.62   -3.2
0  1    0     0.35    -3.2

contrast
0 0 1 -1 0
0 0 -1 1 0


Thanks!
Amanda

########################################################################

To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1

To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1