help
help > RE: Nuisance co-variates
Sep 16, 2015 01:09 PM | Andrew Zalesky
RE: Nuisance co-variates
Hi Patrizia,
It seems like you want to control for an interaction between age and group. In this case, your design matrix should look something like this:
1 1 34 34
1 1 53 53
1 1 23 23
1 0 30 0
1 0 51 0
1 0 22 0
First column is constant, second is the main effect of group, 3rd is the main effect of age and 4th is the age by group interaction. The 4th terms accounts for differences in age between groups.
And your contrast for the main effect of group should be
[0 1 0 0] (patients > controls)
or
[0 -1 0 0] (patients < controls)
The test should be two-sample t-test.
Andrew
Originally posted by Patrizia Dall'Acqua:
It seems like you want to control for an interaction between age and group. In this case, your design matrix should look something like this:
1 1 34 34
1 1 53 53
1 1 23 23
1 0 30 0
1 0 51 0
1 0 22 0
First column is constant, second is the main effect of group, 3rd is the main effect of age and 4th is the age by group interaction. The 4th terms accounts for differences in age between groups.
And your contrast for the main effect of group should be
[0 1 0 0] (patients > controls)
or
[0 -1 0 0] (patients < controls)
The test should be two-sample t-test.
Andrew
Originally posted by Patrizia Dall'Acqua:
Hi NBS community
I am running an analysis were I want to control for age.
My study design has 2 groups (controls and patients) and 2 time points (two-way repeated measure ANCOVA).
I hypothesized that the influence of age on the controls is different from that on the patients. I was wondering, if the design below is correct to control for the influence of age on each group separately.
What about setting up that part of the design matrix which models the covariate in NBS like this (in the example I put 3 subjects per group):
...... Age Patients Age Controls
... 34 0
.... 53 0
.... 23 0
.... 0 30
... 0 51
... 0 22
Technically it works, but I don't know if it is correct.
Thank you in advance for any kind of advice.
Kind regards
Patrizia
I am running an analysis were I want to control for age.
My study design has 2 groups (controls and patients) and 2 time points (two-way repeated measure ANCOVA).
I hypothesized that the influence of age on the controls is different from that on the patients. I was wondering, if the design below is correct to control for the influence of age on each group separately.
What about setting up that part of the design matrix which models the covariate in NBS like this (in the example I put 3 subjects per group):
...... Age Patients Age Controls
... 34 0
.... 53 0
.... 23 0
.... 0 30
... 0 51
... 0 22
Technically it works, but I don't know if it is correct.
Thank you in advance for any kind of advice.
Kind regards
Patrizia
Threaded View
| Title | Author | Date |
|---|---|---|
| Patrizia Dall'Acqua | Sep 14, 2015 | |
| Andrew Zalesky | Sep 16, 2015 | |