Hi Megan,
Perhaps consider setting up your model with 2 2-level within-subject
factors, rather than 1 4-level factor. Ie, a VisVerb factor (Visual or
Verbal) and a RecallRecog factor (Recall or Recognition). The model will
then include main effects for each of those, plus the interactions of
each with Group (and each other). That may make it easier to focus on
the comparisons of interest.
Cheers,
Kylie.
On 3/01/2008 8:04 AM, Megan Malloy wrote:
> I have a study with 2 groups (Group 1 (N=12) and Group 2 (N=10)), all of
> whom took 4 different tests of memory (Test 1= Visual Recall; Test 2= Visual
> Recognition; Test 3= Verbal Recall; Test 4= Verbal Recognition). Using SPSS,
> I ran a repeated-measures ANOVA, Within factor TEST (4 levels) and Between
> factor GROUP (2 levels), and the overall result was significant.
>
> However,
>
> I am interested in answering whether the difference between Group 1 and
> Group 2 on recall is significantly greater than the difference between Group
> 1 and Group 2 on recognition - ?
>
> Also, I am interested in whether the difference between Group 1 and Group 2
> on visual is significantly greater than the difference between Group 1 and
> Group 2 on verbal - ?
>
> What is a valid way to look at these interactions? How do I do post hoc
> testing to get at the answers to these two questions?
>
> Thanks!
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
>
[hidden email] (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD
>
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Locked
