Print

Print
HI Alena,

The earlier email mentioned 2 populations so I guess EV codes for both groups (e.g., +1/-1 or +1/+2). Then yes, -D is necessary, or you can edit the design and add a column full of ones.

All the best,

Anderson


On 24 March 2016 at 19:04, Alena Piovar <[log in to unmask]> wrote:
Dear Anderson,

the design is following 
group  score age
1          231,2   13,3
1         125,4   10,5
1          220,2  12,5
...

contrast

0   1  0
0   -1 0

and I would like to test whether there is a relationship between DTI measures and demean score.

Looking forward to your response.

Best

Alena

2016-03-17 3:24 GMT-05:00 Anderson M. Winkler <[log in to unmask]>:
Hi Alena,

Depends on how the design was coded and if it includes an intercept. Could you show the design?

All the best,

Anderson


On 16 March 2016 at 16:06, Alena Piovar <[log in to unmask]> wrote:
Dear FSL users,

I have a question regarding -D option in randomise is it necessary to adjust it to examine differences between two population adjusted for age as a covariate if the covariate was demeaned by hand (actual values minus mean of age)?

Looking forward to your response

Best

Alena