help > Paired-samples design matrix
Showing 1-3 of 3 posts
Jul 25, 2019 08:07 PM | thomas88
Paired-samples design matrix
Dear NBS community,
I apologize in advance, as I'm sure this is a very straight-forward question, but I am still a little unclear.
I want to run a Baseline vs. Post-intervention connectivity analysis between the same group of subjects, i.e. to see if there are any pre-post changes in connectivity.
If I have, say, 10 subjects, would my design matrix simply be as follows:
0,1
0,1
0,1
0,1
0,1
1,0
1,0
1,0
1,0
1,0
And then would I use the contrasts '[1,-1]' for comparing if pre > post, and '[-1, 1]' for comparing pre < post?
Similarly, if I additionally wanted to compare two separate groups at baseline, could I use the same design matrix for this also (assuming N = 5 for each group)?
Thank you.
I apologize in advance, as I'm sure this is a very straight-forward question, but I am still a little unclear.
I want to run a Baseline vs. Post-intervention connectivity analysis between the same group of subjects, i.e. to see if there are any pre-post changes in connectivity.
If I have, say, 10 subjects, would my design matrix simply be as follows:
0,1
0,1
0,1
0,1
0,1
1,0
1,0
1,0
1,0
1,0
And then would I use the contrasts '[1,-1]' for comparing if pre > post, and '[-1, 1]' for comparing pre < post?
Similarly, if I additionally wanted to compare two separate groups at baseline, could I use the same design matrix for this also (assuming N = 5 for each group)?
Thank you.
Jul 27, 2019 02:07 AM | Andrew Zalesky
RE: Paired-samples design matrix
Hi Thomas,
take a look at the manual for a repeated measures t-test. What you have described here does not account for the fact that the same individuals comprise the baseline and post-intervention group. I.e this is a two-sample t-test, with no repeated measure.
The help button in the GUI also provides an example of a repeated measure t-test.
In short, you design matrix should include a column for each individual and it may be helpful to include an exchange block.
Andrew
Originally posted by Thomas Hill:
take a look at the manual for a repeated measures t-test. What you have described here does not account for the fact that the same individuals comprise the baseline and post-intervention group. I.e this is a two-sample t-test, with no repeated measure.
The help button in the GUI also provides an example of a repeated measure t-test.
In short, you design matrix should include a column for each individual and it may be helpful to include an exchange block.
Andrew
Originally posted by Thomas Hill:
Dear NBS community,
I apologize in advance, as I'm sure this is a very straight-forward question, but I am still a little unclear.
I want to run a Baseline vs. Post-intervention connectivity analysis between the same group of subjects, i.e. to see if there are any pre-post changes in connectivity.
If I have, say, 10 subjects, would my design matrix simply be as follows:
0,1
0,1
0,1
0,1
0,1
1,0
1,0
1,0
1,0
1,0
And then would I use the contrasts '[1,-1]' for comparing if pre > post, and '[-1, 1]' for comparing pre < post?
Similarly, if I additionally wanted to compare two separate groups at baseline, could I use the same design matrix for this also (assuming N = 5 for each group)?
Thank you.
I apologize in advance, as I'm sure this is a very straight-forward question, but I am still a little unclear.
I want to run a Baseline vs. Post-intervention connectivity analysis between the same group of subjects, i.e. to see if there are any pre-post changes in connectivity.
If I have, say, 10 subjects, would my design matrix simply be as follows:
0,1
0,1
0,1
0,1
0,1
1,0
1,0
1,0
1,0
1,0
And then would I use the contrasts '[1,-1]' for comparing if pre > post, and '[-1, 1]' for comparing pre < post?
Similarly, if I additionally wanted to compare two separate groups at baseline, could I use the same design matrix for this also (assuming N = 5 for each group)?
Thank you.
Jul 29, 2019 06:07 PM | thomas88
RE: Paired-samples design matrix
Thanks Andrew, appreciate the response.