Dear Tibor,
 
I am not sure it is ok. If you think that condition A is really just a combination of B and C I think you should use the t-contrast [1 -1 -1].
Another way is to create the first-level contrasts condA>condB+condC (using [1 -1 -1], if this is the correct contrast) and at the second-level do a one sample t test.
The difference between the ANOVA and the t test is the way the errors are considered (pooled against something else I can't remember how it is called now). In my understanding the t-test is more robust.
 
Cheers,
     Laura
 
-----Original Message-----
From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] On Behalf Of Auer, Tibor MD.
Sent: 03 January 2007 19:36
To: [log in to unmask]
Subject: [SPM] Three Conditions, A > B & C

Dear SPMers,

 

At first I would like to wish a Happy New Year for all!

 

I need a help in setting design and contrast to examine the positive effect of conditionA relatively to conditionB and C.

 

I have created the first-level contrast for all the subjects and conditions:

condA > rest,

condB > rest,

condC > rest,

all per 10 subjects.

 

In the next step, I have set up a second-level analysis as a one-way ANOVA with one factor and three levels:

Name: conditions

Level: 3 (A, B, C)

Independence: No (same subjects)

Variance: equal

and with three cells:

1:            10 con*.imgs for condA

                Factors: 1

2:            10 con*.imgs for condB

                Factors: 2

3:            10 con*.imgs for condC

                Factors: 3

 

 

 

To define contrast I used t-contrast, using [2 -1 -1].

 

Does it work for my purpose?

 

Please Help!

 

With many thanks and Best Wishes,

 

Auer, Tibor MD.

PhD-student

Pécs University, Faculty of Medicine

Neurosurgery Clinic

[log in to unmask]

+36-20-986-9341

 



**********************************************************************
This email is confidential and intended solely for the person or entity to whom it is addressed. If this email was not intended for you please notify the UCLH Mail Administrator at [log in to unmask]
This footnote confirms that the email and attachments contained no viruses when they left UCLH.