In general, I would recommend treating sex as a categorical rather than continuous variable. To do this, you would have 6 groups (your 3 x 2 sex). I'm not sure what you are trying to test for. Eg, are you interested in the difference between groups regressing out the effects of age, DD, and sex? Or are you interested in, eg, the effect of age regressing out the other stuff?

On 6/14/2023 3:59 PM, Cristina A. F. Román wrote:

External Email - Use Caution

Hello,

I am looking to run a MANCOVA to examine cortical gray matter differences across three groups (accounting for several covariates), and I was hoping to get input on whether I am setting up my design matrix correctly.

Groups: 3 Covariates: Age (demeaned), disease duration (DD; demeaned), sex

Simplified Example: Group   EV1(Grp1)    EV2(Grp2)     EV3(Grp3)     EV4(age)  EV5(DD)    EV6(sex) 1            1                     0                  0          0.25           0.75            -1 1            1                     0                  0          0.25           0.25            -1 1            0                     1                  0         -0.75          -0.25            1 1            0                     1                  0          0.5            -0.75            1 1            0                     0                  1         -0.5             0.5              1 1            0                     0                  1          0.25           0.25            -1

Questions:

1. Is this an appropriate way to do a three-group comparison with

covariates? Or, is it better to just run three separate t-tests (including covariates)?

2. Regarding post-hoc testing, would I just do three separate t-tests?

This issue here, I imagine, is that the individual t-test results may not reflect the original MANCOVA results (e.g., post-hoc testing reveals a new significant area that was not significant in the MANCOVA).

3. One other option is to run the initial MANCOVA within a statistical

program (i.e., using global cortical gray matter volumes) and then follow up with postdoc testing within FSL (i.e., t-tests). Would this make sense?

Thanks in advance for your time and help, Cristina

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer