Dear Doug

Thank you for the prompt reply.

I want to make sure: If there is no group*age interaction, is it safe to report the result of group differences using DODS (1 -1 0 0 )?

Or is it more appropriate (or mandatory) to further use DOSS to look for group differences (1 -1 0)?

The reason why I ask you is that the results of group differences using DODS is satisfactorily significant to support my hypothesis.

Sorry for bothering you again with my silly question.

Thank you.

Min J.

Date: Tue, 28 Jan 2014 12:10:42 -0500
From: greve@nmr.mgh.harvard.edu
To: freesurfer@nmr.mgh.harvard.edu
Subject: Re: [Freesurfer] A question regarding DODS or DOSS

On 1/28/14 7:03 AM, KimMJ wrote:

Dear Doug and experts

My question is about the choice of DODS or DOSS.

I want to compare cortical thickness between 2 groups (disease group and control group), while controlling for the effect of age (nuisance covariate).

I've already read through the mailing list about the issue of DODS and DOSS designs.

If I understand correctly, I have to first use DODS (QDEC) in order to find areas of significant group*age interaction by using the contrast (0 0 1 -1).

Given there is no area of interaction that survived multiple comparisons correction, I can go with DOSS (1 -1 0) using mri_glmfit (command line) to test between-group differences in cortical thickness, controlling for the effect of age. Am I right?

Yes

My another question is that 'MUST' I rerun DOSS model after confirming no group*age interaction in DODS ? Or can I directly use the results of DODS to look for between-group differences controlling for age (1 -1 0 0), since there is no group*age interaction ?

Is there much difference in results between the above-mentioned two approaches?

Technically, yes, though the results will be different. How different, no one knows. The reason I like going to DODS is that there are no issues with the slopes being slightly different. It does not take much time to re-run it.

I've found a reply by Doug from the mailing list: If there is no interaction, then either DODS or DOSS is appropriate. DOSS will be more powerful and a little more interpretable (http://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg33418.html). From this context, what does the 'powerful' mean? Does DOSS yield higher statistical values than DODS?

It means that you will have a higher degrees of freedom. All other things being equal, it means more significant p values.
doug

Apology for the beginner's questions.

Thank you in advance for your help.

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

_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.