Your comparing the values against 0. Since all your values are likely to be
significantly different than 0, you will have very high F-values.
If your question is where are the groups different, use: [1 -1 0; 0 1 -1]
The two-sample test is still one-sided for a t-test but tests for
differences between groups or average effect of group or where either group
differs from 0 depending on the contrast.
T-test:
H0: group1=group2 --> [1 0 0]*B=[0 1 0]*B --> [1 -1 0]*B=0
HA: group1>group2
F-test:
same as above, except that HA is group1~=group2
Best Regards, Donald McLaren
=================
D.G. McLaren, Ph.D.
Postdoctoral Research Fellow, GRECC, Bedford VA
Research Fellow, Department of Neurology, Massachusetts General Hospital and
Harvard Medical School
Office: (773) 406-2464
=====================
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 (773)
406-2464 or email.
On Mon, Nov 1, 2010 at 2:11 PM, Joel Dunn <[log in to unmask]> wrote:
> Hi,
>
> I'd like to look at Effects of Interest using SPM8 but I am having trouble
> inputting the contrast matrix for the F-contrasts.
>
> The matrices I have tried give me highly significant regions across the
> entire brain: i.e. F-values of at least 1000 in every voxel (within the
> mask).
>
> I'd like to run an equivalent of one-way ANOVA on PET images from 3 groups
> of patients.
> I've created a factorial design - design matrix is 3 columns (n1=14, n2=10,
> n3=12)
>
> I've tried F-contrast matrices 2 different ways both giving the extremely
> high F values
> Using contrast weights matrix = [1 0 0; 0 1 0; 0 0 1]
> OR using the "columns for reduced design" = 0 => [14 0 0; 0 10 0; 0 0 12]
>
> I have tried the same using just 2 groups ("Two Sample T-Test"), assuming
> this is the equivalent of a 2-sided t-test, but the same extreme values
> appear.
>
> Have I entered something incorrectly or made the wrong assumptions?
>
> Cheers
> Joel
> PS - running SPM8 on Matlab R2009a on Windows XP; for simplicity I have
> assumed equal variance and independence between the groups; I used global
> scaling to mean (proportional). Linear T-contrasts show only specific
> regional differences.
>
