The first decision you have to make is whether you want to do pooled or partitioned error. (It's discussed in that manuscript I linked, though admittedly this is all a very technical subject.)
Seems like almost everyone in the SPM community uses pooled error. I'm not really convinced it's a legitimate thing to do, though I seem to be in the minority on that.
If you go with the majority/consensus method of pooled error, a reasonable if somewhat inefficient way to learn how to do things would be to read SPM mailing list posts about it. Searching on "ANOVA" alone might produce too many hits. You could try searching on the phrase "repeated measures" (_with_ double quotes in the search), or
repeated AND measures
One issue is that, at least for your 2x2 design, you have one between- and one within-subject factor. There's no technical problem with that, but it's a little more obscure. I'm pretty sure that Rik Henson once posted a script for such so-called mixed models, but it might be for an older version of SPM.
I'd post more directly on how to do it, but right now my time is limited, so I may or may not.
Best,
S
-----Original Message-----
From: Vaidya, Jatin G [mailto:[log in to unmask]]
Sent: Tuesday, January 05, 2010 10:29 AM
To: Fromm, Stephen (NIH/NIMH) [C]; [log in to unmask]
Subject: RE: [SPM] 2 x 2 ANOVA
Thanks. That makes sense.
I guess I would still like to know how to do this with ANOVA because I want the flexibility of the ANOVA procedure. For instance, if I have a 3x2 design, I'd like to be able to test for all main effects and interaction effects at once.
Jatin
-----Original Message-----
From: SPM (Statistical Parametric Mapping) on behalf of Stephen J. Fromm
Sent: Tue 1/5/2010 7:07 AM
To: [log in to unmask]
Subject: Re: [SPM] 2 x 2 ANOVA
On Mon, 4 Jan 2010 13:17:25 -0500, Fromm, Stephen (NIH/NIMH) [C]
<[log in to unmask]> wrote:
>As far as I know, there _is_ no difference between doing it that way and
doing it as an ANOVA, as long as you're only interested in the interaction term.
Adding, there's no difference in the case you use "partitioned error". There's
probably a difference if you use "pooled error" (see Henson and Penny).
>
>Stephen J. Fromm, PhD
>Contractor, NIMH/MAP
>(301) 451--9265
>________________________________________
>From: Vaidya, Jatin G [[log in to unmask]]
>Sent: Monday, January 04, 2010 10:38 AM
>To: Fromm, Stephen (NIH/NIMH) [C]; [log in to unmask]
>Subject: RE: 2 x 2 ANOVA
>
>Thanks! I know I can run this as a between group t-test after doing the
>subtraction but I'd like to know how to do it as an ANOVA.
>
>The link you sent was helpful. I'm curious if there is any other
>information available on how to actually implement this in SPM (more of
>an applied example).
>
>Thanks for your help.
>
>Jatin
>
>-----Original Message-----
>From: Stephen J. Fromm [mailto:[log in to unmask]]
>Sent: Friday, January 01, 2010 7:53 AM
>To: [log in to unmask]; Vaidya, Jatin G
>Subject: Re: 2 x 2 ANOVA
>
>On Thu, 31 Dec 2009 22:38:19 +0000, Jatin Vaidya <jatin-
>[log in to unmask]> wrote:
>
>>Hi,
>>
>>I'm trying to run a 2 X 2 design with PET data. The first factor is
>group
>(user
>>vs. control). The second factor is task (tapping vs. control); thus,
>this is a
>>repeated measures variable.
>>
>>I can't seem to determine how to run the analysis in SPM with this type
>of
>>dataset. Is there a good resource available for this?
>
>http://www.fil.ion.ucl.ac.uk/~wpenny/publications/rik_anova.pdf
>
>In this particular case, if you're only interested in comparing tapping
>with
>control (i.e., not interested in tapping alone or control alone or their
>average,
>but only their difference), you could also just subtract control from
>tapping at
>the subject level (or using ImCalc), and bring the difference to the
>group
>level, resulting in a 1-way, between-subjects ANOVA with no repeated
>measures.
>
>>
>>Thanks,
>>
>>Jatin
|