Print

Print
Cécilia,

You are correct that you can't add covariates to a two-sample t-test. However, what you have actually performed here is an ANCOVA. If you recall, all of theses analyses (t-test, ANOVA, ANCOVA, simple regression) fall within the construct of the general linear model (GLM). As such, it doesn't matter what you selected in the menus as much as it matters what the final model looks like.

Not sure of any good practical resources except to say that statistics are the same in every program (e.g. do that the same thing you would do in SPSS, SAS, or STATA). Only difference is that you are doing the test on every voxel, instead of just once.

Hope this helps.

Best Regards, Donald McLaren
=================
D.G. McLaren
University of Wisconsin - Madison
Neuroscience Training Program
Office: (608) 520-0586
=====================
This e-mail contains CONFIDENTIAL INFORMATION which may contain PROTECTED HEALTHCARE INFORMATION and may also be LEGALLY PRIVILEGED and which is intended only for the use of the individual or entity named above. If the reader of the e-mail is not the intended recipient or the employee or agent responsible for delivering it to the intended recipient, you are hereby notified that you are in possession of confidential and privileged information. Any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited and may be unlawful. If you have received this e-mail unintentionally, please immediately notify the sender via telephone at (608) 520-0586 or email.


On Thu, Mar 4, 2010 at 1:49 AM, Samieri Cecilia <[log in to unmask]> wrote:

Dear SPM users,
I am writing as a new SPM user to ask a very basic question. I am using SPM5 in VBM to compare two groups of subjects (contrasted according to an exposure variable) and I would like to adjust analyses for age, sex, educational level and TIV. I was told to perform a two-sample t test with adding the four covariates. The contrasts I used were : [1, -1, 0, 0, 0, 0]
I am very surprised that we can add covariates in a ttest, which is not the case in "basic statistics", and I did not find any mention on this possibility in the documentation from Ashburner and coll. However the model seems to be estimated by SPM, which is quite puzzling. In statistics, I would have performed an ANCOVA.
Can anybody confirm (or infirm) me that adding covariates to a two-sample t-test is OK?
Where can I find any practical documentation on how to choose the best model for VBM in SPM ?
Thanks very much by advance for your help,
Cécilia

--


Cécilia SAMIERI, doctorante
Equipe Epidémiologie de la nutrition et des comportements alimentaires Centre de recherche INSERM U897
Université Victor Ségalen Bordeaux 2, ISPED Case 11
146 rue Léo-Saignat 33076 BORDEAUX Cedex - France
Tel: +33 5 57 57 95 38 
Fax: +33 5 57 57 14 86
E-mail : [log in to unmask]